--- language: "en" --- # Knowledge Center ## Documentation ### [Communications Platform](https://knowledgecenter.cm.com/knowledge-center/communications-platform.md) ### [Payments Platform](https://knowledgecenter.cm.com/knowledge-center/payments-platform.md) ### [CM.com Account](https://knowledgecenter.cm.com/knowledge-center/cm-com-account.md) ### [HALO](https://knowledgecenter.cm.com/knowledge-center/halo.md) ### [Mobile Service Cloud](https://knowledgecenter.cm.com/knowledge-center/mobile-service-cloud.md) ### [Mobile Marketing Cloud](https://knowledgecenter.cm.com/knowledge-center/mobile-marketing-cloud.md) ### [Conversational AI Cloud](https://knowledgecenter.cm.com/knowledge-center/conversational-ai-cloud.md) ### [Ticketing](https://knowledgecenter.cm.com/knowledge-center/ticketing.md) ### [Mobile Event App](https://knowledgecenter.cm.com/knowledge-center/mobile-event-app.md) ### [Verification and Sign](https://knowledgecenter.cm.com/knowledge-center/identity-and-sign.md) ### [Central Services](https://knowledgecenter.cm.com/knowledge-center/central-services.md) ### [AI Library](https://knowledgecenter.cm.com/knowledge-center/ai-library.md) * ### [Design dimensions](https://knowledgecenter.cm.com/knowledge-center/design-dimensions.md) * ### [Tasks](https://knowledgecenter.cm.com/knowledge-center/tasks.md) --- language: "en" --- # AI Library * [Mastering AI Prompting](https://knowledgecenter.cm.com/knowledge-center/ai-library/mastering-ai-prompting.md) * [A Beginner's Guide to Artificial Intelligence (AI)](https://knowledgecenter.cm.com/knowledge-center/ai-library/understanding-ai-its-scope-history-and-the-evoluti.md) * [Making Sense of AI: Understanding LLMs, Capabilities, and the Future](https://knowledgecenter.cm.com/knowledge-center/ai-library/navigating-the-overwhelming-world-of-ai-llms-vs-ca.md) * [How Reasoning Models and HALO Work Together to Analyze Complex Documents](https://knowledgecenter.cm.com/knowledge-center/ai-library/how-reasoning-models-and-halo-work-together-to-ana.md) * [When to Use Workflows to Design AI Agents in HALO](https://knowledgecenter.cm.com/knowledge-center/ai-library/when-to-use-workflows-to-design-ai-agents-in-halo.md) * [Preparing to Build Your First Workflow with HALO](https://knowledgecenter.cm.com/knowledge-center/ai-library/preparing-to-build-your-first-workflow-with-halo.md) * [Building and Scaling Workflows with HALO](https://knowledgecenter.cm.com/knowledge-center/ai-library/building-and-scaling-workflows-with-halo.md) * [How to Format AI Prompts for Better Results with HALO](https://knowledgecenter.cm.com/knowledge-center/ai-library/how-to-format-ai-prompts-for-better-results-with-h.md) * [How to Translate the Knowledge Center](https://knowledgecenter.cm.com/knowledge-center/ai-library/how-to-translate-a-web-page-into-your-preferred-la.md) --- language: "en" --- # Building and Scaling Workflows with HALO ## **Introduction** Now that you've prepared your workflow and mapped out your processes, it's time to bring your AI Agent to life with HALO. HALO's agentic AI allows you to design intelligent agents that handle customer interactions seamlessly, whether through email, chat, or phone. This article will guide you through the process of building your AI Agent in HALO and show you how to scale its capabilities to handle more complex tasks. For a practical demonstration of these concepts, including real-world examples, watch the embedded webinar at the end of this article. ## **How to Build an AI Agent in HALO** ### **Step 1: Translate Your Workflow into HALO** Your workflow serves as the blueprint for your AI Agent. Use HALO's interface to define the key components of your agent's behavior: * **Triggers** : What starts the process? *Example*: A customer asks for their shipping status via chat. * **Actions** : What happens next? *Example*: The AI Agent retrieves the tracking number from the CRM and checks the shipping provider's system. * **Outputs** : What is the final result? *Example*: The AI Agent provides the shipping status to the customer in real time. By translating your workflow into HALO, you can ensure your AI Agent is structured to handle tasks efficiently and effectively. ### **Step 2: Use HALO's Agentic AI** HALO's agentic AI is designed to handle complex tasks that require decision-making, context, or integration with external systems. More documentation on HALO's AI Agents and it's possibilities can be found [here](https://knowledgecenter.cm.com/knowledge-center/halo/introduction-to-agents.md). #### **Examples of What HALO's AI Can Do** : * **Summarize Customer Feedback**: Analyze feedback from multiple channels and group it into actionable themes. * **Retrieve Shipping Statuses**: Access external systems to provide real-time updates to customers. * **Analyze Unstructured Data**: Extract key information from emails, chats, or calls. #### **Example Use Case** : A customer asks, *"Where is my package?"* HALO's AI Agent: 1. Ask for a tracking number at the user. 2. Retrieves the shipping status using tool \[TOOL\]. 3. Responds to the customer with the shipping update in a friendly, conversational tone. By leveraging HALO's AI capabilities, you can design agents that go beyond simple automation to deliver intelligent, context-aware responses. ### **Step 3: Test and Refine** Once your AI Agent is built, it's essential to test it thoroughly to ensure it performs as expected. #### **What to Look For During Testing** : * **Errors or Inconsistencies**: Are there any gaps in the agent's responses or actions? * **Simplification Opportunities**: Can any steps be streamlined or removed? * **Automation Potential**: Are there additional tasks the agent could handle? #### **Tips for Refinement** : * Gather feedback from team members or customers who interact with the AI Agent. * Use HALO's analytics tools to monitor the agent's performance and identify areas for improvement. * Iterate on the agent's design to enhance its capabilities over time. ## **Scaling Your AI Agents with HALO** ### **Mixing Deterministic and Probabilistic Steps** As you scale your AI Agent, you may need to combine rule-based (deterministic) steps with AI-driven (probabilistic) decision-making. #### **Example** : * **Deterministic Steps**: Use rule-based automation for structured tasks, such as copying data from one system to another. * **Probabilistic Steps**: Use HALO's AI for tasks that require interpretation or context, such as analyzing customer feedback or generating personalized responses. This hybrid approach ensures your AI Agent is both reliable and flexible, capable of handling a wide range of tasks. ## **Advanced Tips for Building and Scaling AI Agents in HALO** 1. **Write Effective Prompts** HALO's AI relies on clear instructions to perform tasks accurately. When designing your agent, write detailed prompts that specify: * The agent's role (e.g., *"You are a customer support assistant."*). * The task it needs to perform (e.g., *"Retrieve the shipping status for the provided tracking number."*). * The tone or style of the response (e.g., *"Respond in a friendly and professional tone."*). * For more information, check out [this course on AI Prompting](https://knowledgecenter.cm.com/knowledge-center/ai-library/mastering-ai-prompting.md). 2. **Integrate with Systems** HALO's agentic AI can connect to external systems like CRMs, databases, or shipping platforms. This enables your AI Agent to: * Retrieve real-time data (e.g., shipping statuses, order details). * Update records automatically (e.g., logging customer interactions in the CRM). * Provide seamless, end-to-end customer support. ## **Conclusion** Building AI Agents with HALO is a game-changer for automating customer interactions and scaling your operations. By starting with a clear workflow, leveraging HALO's agentic AI, and testing thoroughly, you can create agents that save time, reduce errors, and enhance customer satisfaction. As your confidence grows, you can scale your AI Agents to handle more complex tasks, combining rule-based automation with AI-driven decision-making for maximum flexibility and efficiency. ## **Watch the Webinar** For a practical demonstration of these concepts, watch the webinar below. In this session, we explore how to build and refine an AI Agent, structure workflows, and implement guardrails to ensure your agent performs effectively. --- language: "en" --- # How Reasoning Models and HALO Work Together to Analyze Complex Documents ## Introduction In the fast-paced world of e-commerce, customer service is more than just a support function---it's the backbone of a successful business. Customers expect quick, accurate answers to their questions, whether they're asking about product features, return policies, or troubleshooting steps. But what happens when your business has thousands of pages of user manuals, product descriptions, and internal guidelines? How do you make this wealth of information accessible to both your customers and your team? This is where **HALO**, our Agentic AI solution, comes in. HALO uses advanced reasoning models to analyze complex documents and transform them into actionable knowledge. By doing so, it empowers businesses to deliver exceptional customer service while streamlining internal processes. In this article, we'll explore what reasoning models are, how HALO leverages them, and how this technology can revolutionize your customer service operations. ### What Are Reasoning Models? Reasoning models are a sophisticated type of AI designed to think logically and analyze information in depth. Unlike traditional AI systems that generate quick, surface-level responses, reasoning models excel at understanding context, identifying relationships, and delivering nuanced answers. For example, imagine a customer asks, "How do I reset my device?" A reasoning model doesn't just scan for keywords like "reset" or "device." Instead, it analyzes the entire user manual, identifies the relevant section, and provides a detailed, accurate response tailored to the query. This ability to reason through complex information makes these models ideal for handling large, intricate datasets---like the ones often found in e-commerce businesses. ### How HALO Leverages Reasoning Models HALO isn't a reasoning model itself; rather, it's a platform that integrates reasoning models to make your business knowledge accessible and actionable. The process begins when you upload your documents---such as user manuals, product descriptions, return policies, or internal guidelines---into HALO's [Knowledge](https://knowledgecenter.cm.com/knowledge-center/halo/knowledge.md) section. These documents can be in formats like PDFs or JSON files. Once uploaded, HALO uses reasoning models to analyze the content. The system breaks down the documents into smaller, relevant sections, identifying key information, patterns, and relationships within the text. This analysis transforms static documents into dynamic, searchable knowledge. The real magic happens when this knowledge becomes accessible through HALO's Web Conversations or Mobile Service Cloud. Whether it's a customer asking a question or a service representative seeking information, HALO retrieves and delivers accurate answers in real time. This seamless integration of reasoning models ensures that even the most complex queries are addressed with precision and clarity. ### Real-World Applications in E-Commerce The practical applications of HALO and reasoning models in e-commerce are vast. Consider the following scenarios: #### **1. Simplifying User Manual Queries** Imagine you sell a range of electronic devices, each with a detailed user manual. Customers frequently ask questions like, "How do I update the software?" or "What's the process for resetting my device?" With HALO, you can upload all your user manuals into the system. When a customer asks a question, HALO pinpoints the relevant section of the manual and provides a clear, concise answer---no need for the customer to sift through pages of text. #### **2. Clarifying Return Policies and Warranty Terms** Return policies and warranty terms are often scattered across multiple documents, making it challenging for customers to find the information they need. By uploading these documents into HALO, you enable the system to answer questions like, "How long do I have to return a product?" or "What does the warranty cover?" HALO's reasoning models ensure the answers are accurate and contextually relevant. #### **3. Comparing Products** When customers are deciding between two similar products, they often need detailed comparisons to make an informed choice. HALO can analyze product descriptions and specifications, allowing it to answer questions like, "What's the difference between Product A and Product B?" The result is a detailed comparison that highlights key features and specifications, helping customers make confident purchasing decisions. ### The Benefits of HALO and Reasoning Models The integration of HALO and reasoning models into your customer service strategy offers transformative benefits. Customers enjoy faster, more accurate answers, which leads to higher satisfaction and loyalty. At the same time, your customer service team becomes more efficient, as HALO automates repetitive inquiries, freeing up agents to focus on more complex issues. Even large volumes of information, such as hundreds of pages of user manuals, become easily searchable and accessible. This not only improves the customer experience but also streamlines internal processes, reducing the time and effort required to locate critical information. Additionally, automating common questions can significantly lower the operational costs of customer service, making HALO a cost-effective solution for businesses of all sizes. ### How HALO Handles Large Documents One of the challenges of working with large documents is that AI models can't process every word at once. HALO addresses this using a technique called **Retrieval Augmented Generation (RAG)**. This approach breaks documents into smaller chunks, allowing the AI to search for and analyze the most relevant sections to answer a query. To maximize the effectiveness of HALO, it's important to ensure your documents are well-structured, with clear headings and sections. This organization makes it easier for the AI to locate relevant information. Additionally, asking specific questions---such as "How do I replace the batteries in Model X?" instead of "What does the manual say about batteries?"---can lead to more precise answers. While reasoning models are highly accurate, it's always a good idea to verify responses for complex or sensitive queries to ensure complete accuracy. ### The Future of AI in Customer Service As reasoning models continue to evolve, their potential applications in customer service will only expand. Future advancements could include larger context limits, more sophisticated analytical capabilities, and even the ability to generate personalized recommendations or conduct in-depth customer behavior analyses. For e-commerce businesses, this means even greater opportunities to enhance customer experiences, streamline operations, and stay ahead of the competition. The combination of reasoning models and platforms like HALO represents the future of customer service---one where knowledge is not just accessible but actionable, dynamic, and deeply integrated into every customer interaction. ### Conclusion HALO and reasoning models represent a powerful combination for e-commerce businesses looking to transform their customer service. By uploading your knowledge---whether it's user manuals, return policies, or product descriptions---into HALO, you can provide customers with fast, accurate answers through an intuitive chatbot interface. The result? Happier customers, more efficient operations, and significant cost savings. Ready to take your customer service to the next level? Start uploading your knowledge into HALO today and experience the transformative power of AI-driven reasoning models. --- language: "en" --- # How to Format AI Prompts for Better Results with HALO ## Introduction When working with AI agents like HALO, the quality of your prompts directly impacts the quality of the output. But crafting effective prompts isn't just about what you say; it's also about how you say it. Clear, well-structured prompts are easier to read, interpret, and act upon for both you and the AI. This guide is here to help you get started. By following these practical formatting tips, you'll learn how to create prompts that are not only easy to understand but also optimized for HALO's capabilities. Whether you're a beginner or looking to refine your skills, these techniques will set you up for success. ## Why Formatting Matters Imagine trying to follow a recipe with no headings, no steps, and no clear instructions. Frustrating, right? The same applies to AI prompts. Proper formatting ensures your instructions are clear, logical, and actionable. It reduces ambiguity, minimizes errors, and helps HALO deliver the results you need. Let's explore the essential formatting techniques that will transform your prompts into powerful tools for collaboration with HALO. ## 11 Formatting Tips for Better AI Prompts ### 1. **Use Headings and Subheadings** Headings are your best friend when it comes to organizing information. They make your prompts easy to navigate and help distinguish between different sections. **How to do it:** * Use `##` for main titles. * Use `###` for subtitles. * Use `####` for smaller subtitles. **Why it works:** Clear headings guide both you and the AI, ensuring that each section has a defined purpose. ### 2. **Label Sections Clearly** Labels act as signposts, instantly communicating the purpose of each section. **How to do it:** Combine headings with short, descriptive labels. **Example:** * *Introduction* * *Structure of Your Response* * *Exceptions* * *Important* **Why it works:** Labels eliminate guesswork, making it easier for HALO to interpret your instructions. ### 3. **Use Numbered Lists for Steps** When you need to outline a process or provide step-by-step instructions, numbered lists are the way to go. **How to do it:** Use `1.`, `2.`, `3.`, and so on. **Why it works:** Numbered lists create a logical flow, ensuring HALO follows your instructions in the correct order. ### 4. **Use Bullet Points for Unordered Lists** For guidelines, options, or ideas, bullet points are ideal. **How to do it:** Use `*`, `-`, or `+` to create a bullet point. **Why it works:** Bullet points make information scannable and easy to digest. ### 5. **Use a Slash to Link Internal Actions** HALO allows you to link to internal actions, such as selecting an agent or tool, directly within your prompt. **How to do it:** Use `/` to open a pop-up window and select the desired action. **Why it works:** This feature streamlines workflows and keeps everything connected. ### 6. **Leverage White Space for Readability** A cluttered prompt is hard to read and even harder to interpret. White space is your secret weapon for clarity. **How to do it:** Add a blank line between sections or after a list. **Why it works:** White space reduces visual noise, making your prompts more approachable. ### 7. **Use Horizontal Lines for Visual Breaks** Separators help visually divide sections, making your prompt more structured. **How to do it:** Type `---` on a separate line to create a horizontal rule. **Why it works:** Visual breaks improve readability and help HALO focus on distinct sections. ### 8. **Use Italics for Subtle Emphasis** Sometimes, you need to highlight details without making them too prominent. That's where italics come in. **How to do it:** Place text between `*`. **Example:** *This is an italicized sentence.* **Why it works:** Italics draw attention without overwhelming the reader or the AI. ### 9. **Use Bold Text for Key Emphasis** When something is critical, bold text ensures it stands out. **How to do it:** Place text between `**`. **Example:** **This is a bolded sentence.** **Why it works:** Bold text highlights key points, ensuring HALO prioritizes them. ### 10. **Write Short Sentences** Long, complex sentences can confuse both humans and AI. Keep it simple. **How to do it:** Break information into smaller chunks and avoid unnecessary complexity. **Why it works:** Short sentences are easier for HALO to parse, reducing the chance of errors. ### 11. **Provide Examples for Clarity** Examples are a powerful way to show HALO exactly what you're looking for. **How to do it:** Include a section with example interactions. **Example:** *As an introduction, you could say:* "Hello! I am an AI agent available to assist with any questions you may have. How can I help you?" **Why it works:** Examples eliminate ambiguity and set clear expectations for HALO's responses. ## Final Thoughts Formatting isn't just about aesthetics. It's about functionality. By applying these tips, you'll create prompts that are clear, actionable, and optimized for HALO's capabilities. Remember, the goal is to make your instructions as easy to follow as possible. Whether you're guiding HALO through a complex task or asking a simple question, good formatting ensures you get the results you need. Now it's your turn. Start experimenting with these techniques and see how they improve your interactions with HALO. Happy prompting! --- language: "en" --- # How to Translate the Knowledge Center In today's globalized world, accessing information in your preferred language is crucial for understanding and making the most of the resources available to you. However, not all content is available in every language. For example, our Knowledge Center documentation is currently only available in English, which may not be your preferred language. To help you overcome this challenge, modern browsers offer built-in translation tools that allow you to translate web pages into your desired language quickly and easily. This ensures you can fully understand and utilize the information provided, regardless of the language it was originally written in. Below, we'll guide you through the process of enabling and using translation features in popular browsers, so you can access our Knowledge Center documentation (or any other web page) in the language you're most comfortable with. ## **Why Translate the Knowledge Center Documentation?** Our Knowledge Center is designed to provide you with the information and guidance you need to solve problems, learn new features, and make the most of our products or services. However, we understand that reading technical documentation in a language you're not fluent in can be challenging and time-consuming. By using your browser's translation tools, you can: * **Save Time:** Quickly translate the page into your preferred language without needing external tools. * **Improve Understanding:** Access the content in a language you're fluent in, reducing the risk of misinterpretation. * **Enhance Usability:** Make the Knowledge Center documentation more accessible and user-friendly for everyone. ## **How to Translate Web Pages in Your Browser** ### **1. Google Chrome** Google Chrome has a built-in translation feature powered by Google Translate, making it one of the easiest browsers to use for translating web pages. **Enable Translation in Chrome Settings** 1. Open Chrome and click on the three-dot menu in the top-right corner. 2. Navigate to **Settings** \> **Languages**. 3. Under the "Languages" section, ensure that the **"Offer to translate pages that aren't in a language you read"** option is toggled on. **Translate a Web Page** 1. Open the Knowledge Center documentation or any web page in a language different from your default language. 2. A translation icon will appear in the address bar (shaped like a small document or globe). 3. Click the icon and select your preferred language from the dropdown menu. 4. The page will be translated instantly, allowing you to read the content in your preferred language. ### **2. Microsoft Edge** Microsoft Edge also includes a built-in translation tool powered by Microsoft Translator. **Enable Translation in Edge Settings** 1. Open Edge and click on the three-dot menu in the top-right corner. 2. Go to **Settings** \> **Languages**. 3. Toggle on the option **"Offer to translate pages that aren't in a language I read"**. **Translate a Web Page** 1. Open the Knowledge Center documentation or any web page in a foreign language. 2. A translation prompt will appear at the top of the page. 3. Click **"Translate"** to convert the page into your preferred language. 4. If the prompt doesn't appear, click the translation icon in the address bar. ### **3. Safari (Mac and iOS)** Safari offers translation capabilities for supported languages on macOS and iOS devices. **Enable Translation in Safari** 1. Ensure your Safari browser is updated to the latest version. 2. On macOS, go to **Safari** \> **Preferences** \> **Languages** and add your preferred languages. 3. On iOS, go to **Settings** \> **Safari** \> **Languages** and enable translation. **Translate a Web Page** 1. Open the Knowledge Center documentation or any web page in Safari that is in a foreign language. 2. Click the **Translate** button in the address bar (it looks like a speech bubble with text). 3. Select your preferred language, and the page will be translated. ### **4. Mozilla Firefox** Unlike Chrome and Edge, Firefox does not have a built-in translation tool but supports translation through add-ons. **Install a Translation Add-On** 1. Open Firefox and go to the **Add-ons** page (click the menu button and select **Add-ons and Themes**). 2. Search for a translation extension, such as **To Google Translate** or **Simple Translate**. 3. Install the add-on and follow the instructions to configure it. **Translate a Web Page** 1. Open the Knowledge Center documentation or any web page in a foreign language. 2. Use the add-on's toolbar button or context menu to translate the page into your preferred language. ## **Translating Screenshots and Images** While the browser translation tools described above work excellently for translating text on web pages, they **do not translate text that appears within images or screenshots**. Since our Knowledge Center documentation includes many screenshots showing our product's user interface, buttons, menus, and settings, you may find that these images remain in English even after translating the page. To help you understand the screenshots in your preferred language, you can use **Google Lens**, a free tool that translates text within images in real-time. ### **Using Google Lens in Chrome** Google Lens is built into the Chrome browser, making it easy to translate screenshots without installing additional software. **How to Translate an Image Using Google Lens:** 1. Open the Knowledge Center documentation in **Google Chrome**. 2. **Right-click** on any screenshot or image that contains text you want to translate. 3. Select **"Search image with Google Lens"** from the context menu. 4. A sidebar will open on the right side of your browser showing the Google Lens results. 5. In the Lens sidebar, click the **"Translate"** icon (it looks like a letter "A" with translation symbols). 6. Select your preferred language from the dropdown menu. 7. The text within the image will be highlighted and translated in the sidebar. **Note:** You'll need to repeat this process for each screenshot you want to translate. The translation appears in the sidebar rather than overlaying the original image. ## **Tips for Translating the Knowledge Center Documentation** * **Set Your Preferred Language:** Make sure your browser's default language is set to the one you want translations to appear in. This can usually be done in the browser's language settings. * **Manually Trigger Translation:** If the browser doesn't automatically prompt you to translate, look for the translation icon in the address bar or menu options. * **Bookmark Translated Pages:** Once you've translated a page, bookmark it for quick access in the future. * **Check Context:** Machine translation of screenshots may occasionally be imperfect. Use the translated page text alongside the screenshot translations for better context. ## **Conclusion** We understand that accessing the Knowledge Center documentation in English may not always be ideal for everyone. By using your browser's built-in translation tools, you can easily translate the content into your preferred language, ensuring you have the information you need in a format that works best for you. Whether you're using Chrome, Edge, Safari, or Firefox, the steps outlined above will help you unlock the full potential of our Knowledge Center and make your experience more accessible and user-friendly. If you have any questions or need further assistance, feel free to reach out to our support team. --- language: "en" --- # Mastering AI Prompting Artificial Intelligence (AI) is rapidly transforming the way we work, communicate, and solve problems. At the heart of this transformation lies **AI prompting**---the art of crafting precise instructions to guide AI tools like ChatGPT or HALO. Whether you're a beginner or looking to refine your skills, this guide will walk you through the essentials of AI prompting, helping you unlock the full potential of these tools. ## **What is AI Prompting?** AI prompting is the process of providing clear, structured instructions (called prompts) to an AI model to achieve a desired outcome. Think of it as a conversation with a highly intelligent assistant. The better you communicate, the better the results. A prompt can be as simple as a question or as detailed as a multi-step instruction. The key is clarity and specificity. A vague prompt will yield vague results, while a well-crafted prompt can produce highly accurate and tailored outputs. ## **Why is AI Prompting Important?** Imagine baking cookies without a recipe. You might end up with something edible, but it's unlikely to be the chocolate chip oatmeal cookies you were hoping for. Similarly, prompts act as the "recipe" for your AI. They guide the AI to deliver exactly what you need. By mastering prompting, you can: * Save time by automating repetitive tasks. * Enhance productivity by generating high-quality outputs. * Spark creativity by exploring new ideas and possibilities. ## **The Building Blocks of a Great Prompt** A great prompt has three key components: **Situation** , **Role** , and **Condition**. Let's break these down: 1. **Situation** : Provide the background or context for the task. Example: "I am a business professional researching AI integration." 2. **Role** : Define who the AI should be or what role it should take. Example: "Act as a marketing expert specializing in social media campaigns." 3. **Condition** : Specify the desired output or focus. Example: "Focus on strategies for increasing engagement on Instagram." By combining these elements, you can create prompts that are clear, detailed, and actionable. ## **Examples of Effective Prompts** Let's compare a vague prompt with a detailed one: * **Vague Prompt** : "Tell me about AI." Output: General information about AI. * **Detailed Prompt** : "I am a business professional researching AI integration. Can you provide an overview of how AI can improve decision-making processes in small businesses? Focus specifically on trust and transparency in data usage." Output: A tailored response addressing your specific needs. The difference is clear: better input equals better output. ## **Advanced Prompting Techniques** Once you've mastered the basics, you can explore advanced techniques to further refine your prompts: ### 1. **Step-by-Step Instructions** Break down complex tasks into smaller, manageable steps. Example: * Step 1: Summarize the key points of this article. * Step 2: Rewrite the summary in a conversational tone. * Step 3: Translate the rewritten summary into French. ### 2. **Few-Shot Prompting** Provide examples to guide the AI's output. Example: * Input: "Here are three examples of LinkedIn posts. Use these as a template to create a new post about AI in marketing." ### 3. **Chain-of-Thought Prompting** Encourage the AI to reason step by step. Example: * Input: "First, summarize the meeting notes. Then, draft a follow-up email based on the summary." ### 4. **Self-Consistency Prompting** Simulate multiple perspectives to refine the output. Example: * Input: "Imagine five marketing experts are brainstorming ideas. Provide the top three suggestions they would agree on." ### 5. **Using Boundaries** Clearly define sections of your input to avoid confusion. Example: * Input: "Translate the text between triple quotes into Spanish: `Hello, how are you?`" ## **Common Mistakes to Avoid** 1. **Being Too Vague** : Always provide enough context and detail. Bad: "Write a blog post." Good: "Write a 500-word blog post about the benefits of AI in healthcare, targeting a non-technical audience." 2. **Overloading the Prompt**: Avoid cramming too much information into a single prompt. Break it into steps if needed. 3. **Skipping Examples**: Examples help the AI understand your expectations. Use them whenever possible. 4. **Ignoring Iteration**: Prompting is a trial-and-error process. Refine your prompts based on the AI's responses. ## **Practical Applications of AI Prompting** AI prompting can be applied across various domains, including: * **Content Creation**: Generate blog posts, social media content, or marketing copy. * **Customer Support**: Draft responses to common customer queries. * **Data Analysis**: Summarize reports or extract key insights. * **Education**: Create lesson plans, quizzes, or study guides. * **Personal Productivity**: Automate email drafting, meeting summaries, or task prioritization. ## **The Future of Prompting** As AI continues to evolve, prompting will become an essential skill for professionals across industries. Just as internet literacy became a necessity, prompt engineering will soon be a fundamental part of the modern workplace. ## **Learn More: Video Series on AI Prompting** To help you dive deeper into these themes, we've created a four-part video series. Each video explores one of the key themes in detail, providing actionable insights and practical examples. 1. **Getting Started with AI Prompting** Learn the foundational concepts of AI prompting and why clarity is key. 2. **Mastering the Art of Prompt Engineering** Discover the building blocks of effective prompts and how to guide AI behavior. 3. **Advanced Strategies for Prompt Engineering** Explore advanced techniques for handling complex tasks and edge cases. 4. **Crafting Compelling Prompts for AI Agents** Learn how to structure prompts for HALO's AI Agents and avoid common mistakes. 5. **Designing Your HALO Agent, Not Scripting It** Learn why flexible instructions outperform rigid scripts, and discover the two building blocks every agent starts with. 6. **Guardrails and Handoffs for HALO Agents** Discover how to set firm limits and clean exits, so your agent knows exactly where its job ends. ## **Conclusion: Unlock the Full Potential of AI Prompting** By mastering the techniques outlined in this guide, you'll be able to craft prompts that save time, enhance productivity, and unlock the full potential of AI tools and agents. Whether you're working with ChatGPT, HALO's AI Agents, or any other AI system, these skills will empower you to achieve your goals with precision and efficiency. Start exploring the videos above and put your skills into practice. Happy prompting! --- language: "en" --- # Making Sense of AI: Understanding LLMs, Capabilities, and the Future ## **Introduction** Artificial Intelligence (AI) has become one of the most transformative technologies of our time. It enables machines to perform tasks that once required human intelligence---learning, reasoning, decision-making, and even creativity. From its humble beginnings in the 1930s, when Alan Turing laid the groundwork for machine intelligence, to the sophisticated systems we use today, AI has evolved into a vast and complex field. Modern AI spans across domains like robotics, natural language processing, and machine perception. It powers applications as diverse as fraud detection, autonomous vehicles, and tools that generate human-like text. With the rise of **Large Language Models (LLMs)** like GPT-4, Claude, and Gemini, AI has entered a new era of accessibility and utility. However, the rapid pace of innovation can feel overwhelming. New models, tools, and features emerge almost daily, making it difficult to keep up. The good news? By understanding the core concepts of **LLMs** , **capabilities** , and even the emerging field of **Agentic AI**, you can cut through the noise and focus on what truly matters for your needs. This guide will help you categorize and evaluate AI innovations, empowering you to make informed decisions without feeling lost in the ever-expanding AI landscape. ## **The Building Blocks of AI: LLMs and Capabilities** When a new AI tool or feature is introduced, it typically falls into one of two categories: **Large Language Models (LLMs)** or **capabilities**. These two components work together to define what an AI system can do and how it can help you. ### **LLMs: The Knowledge Base** At the heart of many AI systems lies a **Large Language Model (LLM)**---a powerful engine trained on vast amounts of data to develop knowledge and skills. Think of an LLM as the "brain" of the system. It processes information, generates responses, and performs tasks based on its training. However, not all LLMs are created equal. Each has unique strengths and weaknesses, shaped by the data it was trained on and the tasks it was designed to perform. For example, GPT-4 is a generalist model known for its ability to generate human-like text across a wide range of topics. Claude, on the other hand, excels in conversational tasks, offering user-friendly interactions. Then there's Gemini, a multimodal LLM capable of processing not just text but also images and audio, making it ideal for more complex, cross-domain tasks. An easy way to think about LLMs is to imagine them as new colleagues joining your team. Each colleague brings their own expertise and personality. Some might be great at writing reports, while others excel at analyzing data or brainstorming creative ideas. Understanding the "personality" of an LLM helps you determine how best to use it. ### **Capabilities: The Tools and Skills** While LLMs provide the knowledge base, **capabilities** are the tools that enhance their functionality. These are the specific tasks or functions an AI system can perform, such as searching the web, generating images, or executing code. Capabilities allow you to customize an AI system to meet your needs, making it more versatile and powerful. For instance, an LLM with a search capability can look up real-time information online, while one with a code execution capability can write and run Python scripts. Other capabilities might include generating visuals from text prompts, interpreting PDF documents, or storing and recalling information from previous interactions. Imagine you're working on a marketing project. An AI system with the right capabilities could research market trends, generate ad copy, create visuals, and even analyze campaign performance---all in one workflow. These tools transform an LLM from a static knowledge base into a dynamic assistant that can tackle a wide range of tasks. ## **The Next Frontier: Agentic AI** Beyond LLMs and capabilities lies the emerging field of **Agentic AI**. Unlike traditional AI systems that require user input for every task, Agentic AI operates autonomously. It combines reasoning, planning, and decision-making to achieve user-defined goals without constant supervision. For example, an Agentic AI system tasked with managing a marketing campaign could independently research the target audience, develop strategies, execute tasks, and adjust its approach based on performance data. It integrates multiple capabilities---like search, memory, and code execution---into a cohesive, intelligent agent that can adapt to new information and changing circumstances. Agentic AI represents the next step in AI evolution, offering the potential for systems that can act as true collaborators, not just tools. While still in its early stages, this technology is poised to revolutionize industries by enabling highly autonomous and intelligent systems. ## **How to Evaluate New AI Innovations** With so many new AI tools and features being released, it's important to have a framework for evaluating their relevance. Start by asking two simple questions: 1. **Is it a new LLM?** * Does it bring a new knowledge base or expertise? * Is it trained on different data or designed for a specific purpose? 2. **Is it a new capability?** * Does it add a new function or skill to an existing LLM? * Can it perform tasks that were previously unavailable? For more advanced systems, consider whether it falls under the category of **Agentic AI**. Does it have the autonomy to plan and execute tasks independently? If so, it might offer a level of functionality that goes beyond traditional AI tools. ## **Staying Focused in a Rapidly Changing Landscape** The fast pace of AI development can make it feel like you're constantly playing catch-up. But you don't need to master every new tool or feature to stay ahead. Here are some practical strategies to help you stay focused: * **Adopt a Mental Model:** Think of AI tools as colleagues with different skills. Use the right tool for the right task, just as you would delegate work to team members with specific expertise. * **Prioritize Relevant Capabilities:** Identify the tasks that matter most to your work and focus on tools that excel in those areas. * **Experiment Selectively:** You don't need to try every new tool. Start with those that align with your goals and wait for others to prove their value. * **Leverage Multiple Tools:** Don't rely on a single AI system for everything. Use different tools for writing, coding, image generation, and other tasks as needed. * **Stay Curious:** While you don't need to know everything, maintaining a sense of curiosity will help you adapt to changes and discover valuable innovations. * **Wait for Stability:** Many hype-driven projects fade quickly. Focus on tools and features that demonstrate long-term value and reliability. ## **Conclusion** AI is a vast and constantly evolving field, but by understanding the distinction between **LLMs** , **capabilities** , and **Agentic AI**, you can simplify the landscape. This framework allows you to evaluate new tools, focus on what's relevant, and stay productive without feeling overwhelmed. Remember, you don't need to know everything at once. Take your time, experiment with tools that align with your needs, and trust that the most valuable innovations will stand the test of time. AI is here to assist, not replace, and by leveraging its strengths, you can unlock new levels of efficiency and creativity. For a deeper dive into AI's history, evolution, and the rise of Large Language Models, check out our article [A Beginner's Guide to Artificial Intelligence (AI)](https://knowledgecenter.cm.com/knowledge-center/ai-library/understanding-ai-its-scope-history-and-the-evoluti.md). --- language: "en" --- # Preparing to Build Your First Workflow with HALO ## **Introduction** Imagine an AI Agent that not only answers customer questions but also retrieves real-time data, updates your CRM, and provides personalized recommendations---all while saving you time and resources. With HALO, this is possible, but preparation is the foundation of success. Without a clear plan, you risk creating an AI Agent that is inefficient, incomplete, or unable to meet your business needs. Workflows are the key to ensuring your AI Agent is designed to handle tasks effectively. By mapping out processes and clarifying tasks, workflows act as a blueprint for your AI Agent, helping it deliver real value and integrate seamlessly into your operations. More about this topic, in [this article](https://knowledgecenter.cm.com/knowledge-center/ai-library/when-to-use-workflows-to-design-ai-agents-in-halo.md). Here, we will guide you through the preparation phase, ensuring you're ready to design an AI Agent that meets your business needs. ## **Why Preparation Matters** Think of workflows as the foundation of your AI Agent's success. Without a clear understanding of the steps or inputs required, the result can be chaotic. Workflows help you: * **Identify the right tasks** for your AI Agent to handle. * **Map out the steps** in your process to ensure nothing is missed. * **Leverage HALO's agentic AI capabilities** effectively. Preparation is like onboarding a new colleague. You wouldn't hire someone without providing a clear job description, the right tools, and proper training. Similarly, your AI Agent needs a well-defined role and structure to succeed. By preparing thoroughly, you'll save time, reduce errors, and ensure your AI Agent is set up for success. ## **Step-by-Step Guide to Preparing an AI Agent in HALO** ### **Step 1: Define Your Use Case** Start by identifying a small, simple task that your AI Agent can handle. Focus on tasks that are repetitive, time-consuming, or prone to errors. **Example**: Retrieving the shipping status of a package for customers. **Ask Yourself:** * **What triggers the task?** (e.g., a customer asks for their shipping status via chat.) * **What is the desired outcome?** (e.g., the AI Agent retrieves the shipping status and provides it to the customer.) * **How often is the task performed?** (e.g., multiple times a day.) Starting with manageable tasks like retrieving shipping statuses allows you to test HALO's capabilities while minimizing complexity. This approach builds confidence and sets the stage for scaling to more complex workflows. ### **Step 2: Map Out Your Current Workflow** Before designing your AI Agent, map out the current process for handling the task. This will help you identify the steps your AI Agent needs to replicate or improve. #### **Tips for Mapping Your Workflow:** * Write each step on a separate post-it or use a digital tool like **Lucidchart** or **Miro** to create a flowchart. * Include **inputs** (what triggers the step) and **outputs** (what happens next). * Identify repetitive or time-consuming steps that could benefit from automation. #### **Example Workflow for Retrieving Shipping Status:** 1. A customer asks for their shipping status via chat. 2. The agent retrieves the tracking number from the CRM. 3. The agent checks the shipping provider's system for the status. 4. The agent provides the shipping status to the customer. Workflows act as a "blueprint" for your AI Agent. By clearly mapping out each step, you ensure your AI Agent performs as expected and integrates seamlessly into your existing processes. ### **Step 3: Evaluate the Role of HALO's AI** Not every step in your workflow requires AI. Some tasks are better handled with simple automation, while others benefit from HALO's agentic AI capabilities. #### **When to Use HALO's AI:** * For **unstructured data**, like interpreting customer queries or extracting information from emails. * For **decision-making**, like determining the best response based on context. * For tasks requiring **integration with multiple systems**, like retrieving shipping statuses from external platforms. #### **When to Use Simple Automation:** * For **structured tasks**, like copying data from one system to another. * For **straightforward actions**, like sending a confirmation email. HALO's agentic AI is particularly powerful for tasks that require reasoning, context, or dynamic decision-making. For example, an AI Agent in HALO can retrieve shipping statuses from external systems and provide personalized updates to customers. ## **Examples of Workflows for AI Agents in HALO** Here are some real-world examples of tasks you can structure with workflows to design effective AI Agents in HALO: 1. **Shipping Status Updates** : Design an AI Agent that retrieves the shipping status of a package from external systems and provides the customer with real-time updates via chat, email, or phone. 2. **CRM Updates** : Create an AI Agent that extracts customer details from emails or chats and updates your CRM system automatically. 3. **Feedback Categorization** : Build an AI Agent that analyzes customer feedback from multiple channels and groups it into themes for actionable insights. 4. **Product Recommendations** : An AI Agent can analyze customer preferences and recommend products based on past purchases or browsing history. 5. **Return Management** : An AI Agent can guide customers through the return process, including generating return labels and updating inventory systems. ## **Tips for Getting Started with HALO** 1. **Start Small** Begin with a simple use case to test HALO's capabilities. For example, design an AI Agent to handle shipping status inquiries before expanding to more complex tasks. 2. **Use Tools in HALO** HALO provides tools to define actions and outputs for your AI Agent. Use these to structure your agent's behavior step by step. 3. **Leverage Resources** HALO's Knowledge Center offers tutorials, best practices, and examples of how to design effective AI Agents. These resources can help you refine your workflows and maximize HALO's potential. 4. **Iterate and Improve** Once your AI Agent is live, gather feedback and refine its behavior. Start with a basic version and gradually add more capabilities as you gain confidence. ## **Common Challenges and Solutions** * **Challenge** : *"What if my workflow is too complex?"* **Solution**: Start with a smaller, simpler task and expand as you gain confidence. HALO's tools make it easy to iterate and improve. * **Challenge** : *"How do I integrate HALO with my existing systems?"* **Solution**: HALO supports API integrations with platforms like Shopify and Salesforce, making it easy to connect your systems. * **Challenge** : *"How do I monitor my AI Agent's performance?"* **Solution**: Use HALO's monitoring tools to review conversations, analyze agent performance, and provide feedback for continuous improvement. ## **Conclusion** Preparation is the foundation of a successful AI Agent in HALO. By defining your use case, mapping your workflow, and evaluating HALO's role, you'll be ready to design an AI Agent that works seamlessly and delivers real value. To explore these concepts further, watch the webinar below, where our AI specialists discuss practical examples, best practices, and strategies for building effective AI Agents with HALO. --- language: "en" --- # A Beginner’s Guide to Artificial Intelligence (AI) ## Introduction Welcome to our comprehensive guide on Artificial Intelligence (AI)! Whether you're a curious beginner or someone looking to integrate AI into your daily tasks, this guide will help you understand the fundamentals of AI, its history, and its practical applications. By the end, you'll have a solid grasp of what AI is, how it works, and how it's shaping the world around us. Let's dive in! Don't forget to check out the video below for a deeper dive! ## What is Artificial Intelligence? Artificial Intelligence, or AI, refers to the ability of machines to perform tasks that typically require human intelligence. These tasks range from recognizing patterns and making decisions to generating creative content and holding conversations. AI is already a part of our everyday lives, often in ways we don't even notice. For example, when Spotify curates a playlist based on your listening habits, that's AI at work. When your bank flags a suspicious transaction, AI is analyzing patterns to detect fraud. Tesla's self-driving cars? They rely on AI to make split-second decisions on the road. Even your smart thermostat, like Nest, uses AI to learn your preferences and optimize energy usage. And, of course, tools like ChatGPT use AI to engage in human-like conversations. In short, AI is all about making machines smarter so they can assist, enhance, or even automate tasks that were once the sole domain of humans. ## A Brief History of AI AI may seem like a modern marvel, but its roots stretch back nearly a century. The journey began in 1936 when Alan Turing introduced the concept of a machine capable of simulating human intelligence. This idea laid the foundation for what we now call the Turing Test, a benchmark for determining whether a machine can exhibit human-like behavior. Over the decades, AI has evolved through several key milestones: * **Machine Learning (ML):** This approach involves training models with labeled data to make predictions or decisions. For example, an ML model might analyze thousands of images to learn how to identify cats in photos. * **Deep Learning:** A subset of ML, deep learning uses neural networks to automatically learn features from data. This allows models to recognize complex patterns, such as distinguishing between different accents in speech or identifying objects in videos. * **Generative AI:** The latest frontier in AI, generative models use advanced techniques like transformers to understand context and create new content. This includes everything from generating realistic images to writing coherent text, as seen in tools like ChatGPT. Each of these milestones has brought us closer to the sophisticated AI systems we use today. ## Types of AI Models AI models can be broadly categorized based on their capabilities and applications. Here's a closer look at the three primary types: 1. **Traditional AI:** These models are task-specific and excel at solving narrowly defined problems. For instance, fraud detection systems in banks are trained to analyze transaction data and flag anomalies. Traditional AI is highly effective for repetitive, rule-based tasks but lacks the flexibility to adapt beyond its training. 2. **Generative AI:** Unlike traditional AI, generative models are pre-trained on massive datasets and can create new content. They're used in applications like text generation, image synthesis, and even music composition. Generative AI has gained significant attention for its ability to produce human-like outputs, making it a game-changer in fields like marketing, customer service, and creative industries. 3. **Agentic AI:** This emerging type of AI goes beyond task-specific or content-generating capabilities. Agentic AI combines reasoning, planning, and decision-making to autonomously achieve user-defined goals. These systems can interact with their environment, adapt to new information, and execute complex tasks without constant human intervention. For example, an AI agent might autonomously plan and execute a marketing campaign, adjusting strategies in real time based on performance data. Agentic AI represents the next frontier in AI development, offering the potential for highly autonomous and intelligent systems. ## Layers of AI: Machine Learning, Deep Learning, and Beyond To better understand AI, it's helpful to break it down into its core layers: ### Machine Learning vs. Deep Learning * **Machine Learning (ML):** Think of ML as the foundation of AI. It involves training models with structured data and predefined features. For example, an ML model might be trained to predict house prices based on factors like location, size, and number of bedrooms. * **Deep Learning:** Deep learning takes things a step further by using neural networks to automatically learn features from raw data. This makes it especially powerful for tasks like image recognition, where the model can identify patterns without human intervention. ### Generative AI and Large Language Models Generative AI, a subset of deep learning, uses advanced architectures like transformers to understand context and generate new content. Large language models (LLMs), such as GPT, are trained on vast datasets to predict the next word or token in a sequence. This enables them to power conversational interfaces, write articles, and even assist with coding tasks. ## Best Practices for Using AI To make the most of AI, it's important to approach it strategically. Here are some tips to guide you: * **Choose the Right Model:** Depending on your goals, you might need a non-reasoning model for simple tasks, a reasoning model for complex decision-making, or even a multimodal model that combines text, images, and other data types. * **Leverage AI Tools:** Enhance your AI capabilities with tools like real-time search, API integrations, and pre-built models. These tools can save time and make AI more accessible. * **Experiment with AI Agents:** AI agents are advanced systems that combine reasoning, planning, and tool execution to autonomously achieve user-defined goals. They're ideal for tasks that require a high degree of autonomy and adaptability. By understanding these best practices, you can unlock the full potential of AI and apply it effectively in your work or personal projects. ## Conclusion Artificial Intelligence is no longer a futuristic concept---it's here, and it's transforming the way we live and work. From its humble beginnings with Alan Turing to the sophisticated generative models of today, AI has come a long way. By understanding its history, types, and layers, you'll be better equipped to harness its power for your own needs. For a deeper dive into AI concepts and practical applications, don't miss the video session below. It covers everything from foundational definitions to advanced model types, offering valuable insights for beginners and experts alike. ## Watch the Video Session Ready to take your AI knowledge to the next level? Watch our in-depth video session, where we explore AI concepts, history, and real-world applications in greater detail. Click the link below to get started. --- language: "en" --- # When to Use Workflows to Design AI Agents in HALO ## **Introduction** In today's fast-paced world, repetitive tasks can drain time and energy. That's where workflows come in. Workflows are a powerful way to map out processes and structure tasks, helping you design effective AI Agents in HALO. With HALO's agentic AI capability, you can streamline customer interactions, reduce errors, and free up time for more meaningful conversations. But how do you know when a workflow is the right solution to help design an AI Agent? This article will help you identify the tasks that are perfect for automation and show you how workflows can guide the design of your AI Agents in HALO. ## **What Are Workflows and How Do They Help in HALO?** Workflows are step-by-step representations of processes that help you organize and structure tasks before building AI Agents in HALO. They allow you to break down complex processes into manageable steps, ensuring your AI Agent is designed to handle tasks efficiently and effectively. Unlike traditional automation tools, HALO's agentic AI can handle both structured and unstructured data, making it ideal for tasks that require decision-making or context. * **Structured data** refers to information that is organized and formatted in a predefined way, such as rows and columns in a database or fields in a CRM system. For example, retrieving a customer's order number and shipping status from a database is a task that involves structured data. * **Unstructured data**, on the other hand, is information that doesn't follow a specific format, such as free-text customer feedback, email content, or chat transcripts. For instance, analyzing a customer's email to extract their complaint and categorize it into themes (e.g., "delivery delay" or "product defect") involves unstructured data. **Example** : Imagine a customer sends an email saying, *"Hi, I ordered a laptop last week, but I haven't received it yet. Can you check the status for me?"* * A traditional automation tool might struggle to interpret this unstructured text and extract the relevant details (e.g., "laptop," "last week," "check status"). * HALO's agentic AI, however, can analyze the email, identify the intent (shipping inquiry), extract key details (product: laptop, timeframe: last week), and retrieve the shipping status from an external system. It can then respond to the customer with a clear update, such as: *"Your laptop is scheduled for delivery tomorrow."* By using workflows to design your AI Agent, you can ensure it is equipped to handle both structured tasks (like retrieving data from a database) and unstructured tasks (like interpreting free-text queries) seamlessly. For example, workflows can help you design AI Agents in HALO to: * Extract key information from emails, chats, or calls. * Retrieve shipping statuses or order details from external systems to assist customers in real time. * Summarize customer feedback into actionable insights. * Update CRM systems with customer data. ## Why Use Workflows to Design AI Agents? Workflows are invaluable when designing AI Agents because they bring clarity, scalability, and consistency to the process. By mapping out tasks step by step, workflows help you identify what your AI Agent needs to do and how it should handle each interaction. This structured approach minimizes errors and ensures your AI Agent delivers accurate, reliable results. Additionally, workflows make it easier to scale your AI Agent's capabilities as your business grows. Whether you're handling an increasing volume of customer inquiries or expanding into new use cases, workflows provide a solid foundation for building AI Agents that can adapt to your needs. By using workflows, you can save time and resources during the design process, focus on higher-value activities like improving customer satisfaction, and ensure your AI Agent consistently delivers exceptional results. ## When to Use Workflows in HALO Not every task requires a workflow, so how do you know when it's the right approach? Workflows are particularly useful for tasks that are repetitive, scalable, or require context and consistency. For example, if a task is performed multiple times a day---such as retrieving shipping statuses for customers---it's a strong candidate for a workflow. Similarly, if the task volume is increasing and manual handling is no longer feasible, a workflow can help you automate the process and ensure scalability. Tasks that involve unstructured data, like analyzing customer feedback or interpreting free-text queries, also benefit from workflows. By structuring these tasks clearly, you can ensure your AI Agent handles them accurately and consistently. ## **Examples of Workflows for AI Agents in HALO** Here are some real-world examples of tasks you can structure with workflows to design effective AI Agents in HALO: 1. **Shipping Status Updates**: Design an AI Agent that retrieves the shipping status of a package from external systems and provides the customer with real-time updates via chat, email, or phone. 2. **CRM Updates**: Create an AI Agent that extracts customer details from emails or chats and updates your CRM system automatically. 3. **Feedback Categorization**: Build an AI Agent that analyzes customer feedback from multiple channels and groups it into themes for actionable insights. ## **Tips for Evaluating Use Cases** Before designing an AI Agent in HALO, use workflows to evaluate the task: * **What is the task?** Clearly define the process you want the AI Agent to handle. * **What is the desired outcome?** Identify the result you want the AI Agent to achieve. * **Can HALO handle it?** Determine if HALO's agentic AI can manage the task's complexity, such as retrieving data from external systems or interpreting unstructured information. By starting with small, manageable tasks, you can build confidence and gradually scale your AI Agent's capabilities. ## **Conclusion** Workflows are an essential tool for designing effective AI Agents in HALO. By mapping out processes, clarifying tasks, and identifying the right use cases, you can ensure your AI Agent is equipped to handle customer interactions seamlessly. Ready to get started? Check out our next article on [preparing to build your first AI Agent with HALO](https://knowledgecenter.cm.com/knowledge-center/ai-library/preparing-to-build-your-first-workflow-with-halo.md). --- language: "en" --- # Central Services * [Analytics Portal](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal.md) * [Conversational Router](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router.md) * [Web Conversations](https://knowledgecenter.cm.com/knowledge-center/central-services/web-conversations-documentation.md) * [Marketplace](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace.md) --- language: "en" --- # Analytics Portal * [General information Analytics Portal](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/general-information-analytics-portal.md) * [Tools and Features](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools.md) * [Settings in the Analytics Portal](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/settings-in-the-analytics-portal.md) * [FAQ](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/faq.md) --- language: "en" --- # FAQ * [I cannot login to the Agent Inbox / Mobile Service Cloud (FAQ)](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/faq/i-cannot-login-to-the-agent-inbox-mobile-service-1.md) * [Where can I find the portal?](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/faq/where-can-i-find-the-portal.md) * [Who can access the portal?](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/faq/who-can-access-the-portal.md) * [Where can I find the latest updates?](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/faq/where-can-i-find-the-latest-updates.md) * [How can I provide feedback?](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/faq/how-can-i-provide-feedback.md) --- language: "en" --- # How can I provide feedback? **Idea Portal** To provide feedback or report an issue on the analytics portal, dashboards, insights, features, or other improvements, click on 'Idea Portal' in the bottom left corner to access the CM.com Engage Feedback Portal. Alternatively, visit [++**this link**++](http://support.cm.com/)directly. We'd love to hear from you! ![c5351f0a-e4eb-4020-92f9-7c31be9d683f-image.png](https://knowledgecenter.cm.com/__attachments/a_c249569633757caac126ab0b4bf10a558395fb8a4f849925b5d1b414f3f65b16/c5351f0a-e4eb-4020-92f9-7c31be9d683f-image.png?cb=233b48d55f20bb7ef965ef58f6a058e7) --- language: "en" --- # I cannot login to the Agent Inbox / Mobile Service Cloud (FAQ) ## Issue When logging into MSC you get a message you can't log in. The message will look like this: ![Login_failed.png](https://knowledgecenter.cm.com/__attachments/a_ad5ea30bfa5969fcb4a744de7bd488d27189c2a5593698f3f0d273ad26bb9e66/Login_failed.png?cb=7353b2e4dd666e73703c777a3b163581) ## Product Mobile Service Cloud ## Resolution Please check the entered e-mail address and password and try again. If you are still unable to log in, please select: "Forgot your password?". You will be able to enter the email address you use to log in. You will receive an e-mail to reset your password. ![Schermafbeelding_2022-09-30_154833.png](https://knowledgecenter.cm.com/__attachments/a_9cc93800313e8306b02ba0d07247442c80699f7e660fb154865b292172f5faa0/Schermafbeelding_2022-09-30_154833.png?cb=e0539beab0bf57a12d44aabfc3c8bf0a) ## I don't have access to the inbox of the e-mail address I use to log in: If you do not have access to the inbox of the e-mail address you use to log in, please get in contact with someone within your organization who has access to the settings in MSC. They will be able to change your password. They will be able to change the password by navigating to the Settings ⇾ Team ⇾ \[Your profile\]. Here, you will be able to enter a new password. Make sure you save this new password at the top of the page. ![Password_profile_settings.png](https://knowledgecenter.cm.com/__attachments/a_992efcfc53778e96b9bd6f97c9682a8bd925cfa3ec1b50af4afa4ecbfe5c4e48/Password_profile_settings.png?cb=b4f0925f1479b28e4aab85e5c1e640ae) ## My profile has been deleted, can I get it back? When you can not find the profile you are looking for, it may be deleted. Please contact support@robinhq.com. The support representative will be able to reactivate the account. If you have access to the command center, go to user management and set the deleted user's account back to its original account type. (Regular/Collaboration). This will automatically restore the account to it's latest state. ## Background The e-mail address used to log in does not have to be a real/live e-mail address. This can be fictional. We do not recommend to use fictional e-mail addresses for log in. You will not be able to receive the mail to reset your password. --- language: "en" --- # Where can I find the latest updates? ## **What's new** To stay up-to-date with the latest additions to the Analytics Portal app and its dashboards, click on 'What's New' in the bottom left corner for a direct link to the release notes page. You can also visit the release notes directly at [++**this link**++](https://releasenotes.cm.com/rnotes/analytics-portal). ![Screenshot_2024-11-13_at_12.03.19_copy.png](https://knowledgecenter.cm.com/__attachments/a_bb3fd3aa9c1f51122383040026be28c96dd8ad5e380fc1d57253c72b6b840b3d/Screenshot_2024-11-13_at_12.03.19_copy.png?cb=d67569a5ca85bb238ed431d25ddb09b3) --- language: "en" --- # Where can I find the portal? The Analytics Portal is accessible via the 9-tile menu (in the central services section) for all Mobile Service Cloud Pro customers. ![Screenshot_2024-10-07_at_13.09.59.png](https://knowledgecenter.cm.com/__attachments/a_c59c0ceaf681c0a2ef86d0dc52157d0552c4fde9689f4b171f44f308291d9c94/Screenshot_2024-10-07_at_13.09.59.png?cb=c8a3d041b7a2379680686409fce0b3d8) Please watch the video below on how to log in to the Analytics Portal. It also provides an overview of the Analytics Portal app. --- language: "en" --- # Who can access the portal? The analytics portal is accessible for all users within the logical account. However, dashboard access is limited to certain platform roles. * Users with the ***'Platform owners'*** role, have access to dashboards and data across logical accounts. * Users with an ***'Administrator'*** role have access to dashboards and data within their allocated logical accounts * Users with the ***'User'*** role can reach the portal, but cannot view any of the available dashboards. --- language: "en" --- # General information Analytics Portal * [Introduction to the Analytics Portal](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/general-information-analytics-portal/introduction-to-the-analytics-portal.md) * [The Overview](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/general-information-analytics-portal/the-overview-page.md) * [Available reports](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/general-information-analytics-portal/available-reports.md) --- language: "en" --- # Available reports ## Reports By clicking on one of the available reports on the overview page, you will be directed to the first dashboard within that product report. The analytics portal currently provides analytics reports for the following products: * [Mobile Service Cloud](https://knowledgecenter.cm.com/knowledge-center/mobile-service-cloud/data-and-insights/analytics-dashboards/what-advanced-analytics-dashboarding-is-available.md) * Cash register (POS) * Ticketing (YTP) * Ticketing (M\&P) --- language: "en" --- # Introduction to the Analytics Portal The Analytics Portal is a CM.com platform tool that centralizes all your advanced product analytics into a single application. It allows you to analyze and optimize the usage of CM.com products to better achieve your organizational goals. To use the Analytics Portal, one of the supported product tiers needs to be acquired to enable it. ![Screenshot_2024-11-11_at_17.28.26.png](https://knowledgecenter.cm.com/__attachments/a_3703e47237e91ea4cbc4e28d24ef886c59c64c1c88b7bc00ab8e1004ebe24e49/Screenshot_2024-11-11_at_17.28.26.png?cb=febaf8c10f023484e0780de355febec8) --- language: "en" --- # The Overview The Overview Page serves as your analytics portal hub, displaying the reports available to you within your organisation. ![Screenshot_2024-11-13_at_12.08.09.png](https://knowledgecenter.cm.com/__attachments/a_d582d7c622f29a5b044bcea2ec6a3970c9589b279751dcb56008bfd20f9e2a59/Screenshot_2024-11-13_at_12.08.09.png?cb=15e6aa2048461e679be3ba13ad3d4709) Tip: If you know your organization has access to a particular report and you do not see it, this might be due to your user rights. For more information, see [Who Can Access the Portal?](cards/6f4d0a3c-236f-4951-8ed0-f0ce9a4ae90b) ## Reports By clicking on one of the available reports, you will be directed to the first dashboard within that product report. The analytics portal currently provides analytics reports for the following products: * [Mobile Service Cloud](https://knowledgecenter.cm.com/knowledge-center/mobile-service-cloud/data-and-insights/analytics-dashboards/what-advanced-analytics-dashboarding-is-available.md) * Cash register (POS) * Ticketing (YTP) * Ticketing (M\&P) --- language: "en" --- # Settings in the Analytics Portal * [Users settings](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/settings-in-the-analytics-portal/users-settings.md) --- language: "en" --- # Users settings ## Introduction The user settings page allows you to view and alter analytics portal user roles. User roles define acces to settings and analytics. ### User overview When you land on the users settings page, you are presented with an overview of all users available within your platform account. If a user is missing, you can invite this user via the "Invite user(s)" button, that will redirect you to the platform account, users page. Once a user is added to the platform account, it will become available within the Analytics Portal users settings page. ### User Roles User roles define acces to settings and analytics. Two roles may be allocated to users that define a users acces priviliges. **Analyst** The analyst role can be allocated to regular platform users. By allocating the analyst role, you can provide access to analytics reports, without providing their admin or platform owner privileges within the platform account. **Admin** The admin role is automatically allocated to admin and platform owner users within the platform account. Regular users may also be allocated this role to provide them acces to the available analytics reports, as well as access to the settings. ![Screenshot_2025-01-27_at_16.15.17.png](https://knowledgecenter.cm.com/__attachments/a_7ffce41eb4bb3ad4f378485edb5515562667e1be5b80caae2553a8b3ebed8a4b/Screenshot_2025-01-27_at_16.15.17.png?cb=9df9a9a6896f56435f2db509a974720f) --- language: "en" --- # Tools and Features * [Introduction to the tools](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools/introduction-to-the-tools.md) * [Dashboard Filtering](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools/dashboard-filtering.md) * [Filterset Bookmarking](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools/filterset-bookmarking.md) * [Subscriptions](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools/subscriptions.md) * [Data Exporting](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools/data-exporting.md) --- language: "en" --- # Dashboard Filtering In order to find and analyze the metrics of interest, the dashboards offer extensive filter options that allow you to drill down into the specific insights you are looking for. Filters come in various forms; the following sections describe the types available and how you can use them: ## Panel filters Some dashboards support panel filters, which can be used to filter data across the entire dashboard, affecting all widgets within it. If panel filters are available, they can be found under the filter button at the top right of the dashboard. ![Screenshot_2024-11-13_at_10.46.26.png](https://knowledgecenter.cm.com/__attachments/a_02f509cce3e96685977d7a5807c445094eff53ba09f70606627ab01b2c7d9c02/Screenshot_2024-11-13_at_10.46.26.png?cb=d3c24f25ce60e1bba6d6367e30121a76) **Relative Period Slicers** The filter pane offers various slicers to help filter the dashboard data. A relative period selector within the filter pane allows you to easily choose the period scope of interest. You can quickly select a relative period using the period selector buttons, or choose a specific timeframe using the slider. ![Screenshot_2024-11-13_at_10.45.38.png](https://knowledgecenter.cm.com/__attachments/a_14844fd00e58998ffb0bff91811aa7df48c437e4a113dc813d1b17095319a15e/Screenshot_2024-11-13_at_10.45.38.png?cb=87848b99ece61781ccb79dce58144d1c) **Parameter Slicers** Within the filter pane, parameter slicers may be available. A parameter slicer allows you to select one or multiple values as filters, such as choosing which metric to display or selecting the number of webstores to specify the dashboard's scope. ![Screenshot_2024-11-13_at_10.46.20.png](https://knowledgecenter.cm.com/__attachments/a_506aeec8735a6b1f464bc1e473e17f744f1c79660f81349ab081aa2f5a61d24c/Screenshot_2024-11-13_at_10.46.20.png?cb=e49ac6ede4600021e72e4a3bd7971aeb) ### Widget filters Some dashboards support widget filters, which can be used to filter data for a specific widget on a dashboard. These widget filters can be combined with panel filters for more comprehensive data analysis. **Parameter Slicers** For widgets, parameter slicers may also be available. A parameter slicer allows you to select one or multiple values as filters, such as choosing which metric to display or the period granularity of a trend. ![Screenshot_2024-11-13_at_10.45.20.png](https://knowledgecenter.cm.com/__attachments/a_88b76d0d353f8f5c39ebfbfdcd2454390675a0d33b4d2b00c9c1423979381913/Screenshot_2024-11-13_at_10.45.20.png?cb=3b52ba64c75773d353e7d8fb1bee7942) ### Explainer video The following video provides clear explanation of how to filters within the Analytics Portal dashboards. Explore these resources to improve your workflow and make the most of your Analytics Portal! --- language: "en" --- # Data Exporting To extract data from dashboards or widgets, you can export it into CSV and Excel file formats. For a brief guide on how to use this feature, please watch the instructional video provided below. --- language: "en" --- # Filterset Bookmarking To simplify saving and revisiting your most valued dashboards with specific filter sets, we've added a Bookmarks feature. This allows you to revisit your favorite dashboard filters without manually resetting them each time. Bookmarks are saved at the organizational level, so all users can access them. Interested in using this feature? The following video clearly explains how to utilize bookmarks within the Analytics Portal. --- language: "en" --- # Introduction to the tools At launch, the Analytics Portal supports various tools to help you navigate and extract valuable insights from your data. The following sections describe these tools. --- language: "en" --- # Subscriptions In order to be pro-actively inform platform users about the latest performance insights of your organisation, it is possible to create subscriptions to bookmarks. The Subscribe feature allows you to pro-actively and recurrently receive your favorite dashboard insights straight into your e-mail box (without manually having to search for them on the platform). Subscriptions are saved at the organisational level, and can be sent to multiple recipients at the same time. ## **Creating a subscription** In order to receive dashboard insights into your e-mail inbox, you can subscribe to a bookmark that represents a specific dashboard (including its bookmarked set of filters). So before you are able to create a subscription, first create a bookmark of the dashboard you would like to recieve in your e-mail inbox. For a guide, please find: [Filterset Bookmarking](https://knowledgecenter.cm.com/knowledge-center/central-services/analytics-portal/tools/filterset-bookmarking.md) ### **Subscribing to a bookmark** To subscribe to one of your created bookmarks, you have to navigate to the bookmarks button in the top left corner of the screen. ![Screenshot 2025-05-21 at 14.52.45.png](https://knowledgecenter.cm.com/__attachments/a_d8d98c19eada4becb61b735da5d85c68e260d24eb5b735eefdcf5afed8e99dda/Screenshot%202025-05-21%20at%2014.52.45.png?cb=58e69adf6a19474b2aaf4d8b351aa292) Once you have opened the bookmarks overlay on the right side of the screen, you have to select the three dots icon, in order to find the option to "Subscribe" a bookmark. ![Screenshot 2025-05-21 at 10.21.09-20250521-082115.png](https://knowledgecenter.cm.com/__attachments/a_c395e49d8cc08fb9bbec91d8d1786041846636e9bf8e2a3741a706442b732909/Screenshot%202025-05-21%20at%2010.21.09-20250521-082115.png?cb=69154208c2597ad279b0baabb657674c) ### **Configuring the subscription** When the "Subscribe" button is pressed, the "Create Subscription" options become available. On this screen you provide the following information: ***Start date:*** Date from which the subscription is first sent to the recipient(s). ***End date (not mandatory):*** Date at which the subscription ends, and thus is no longer received by the recipient(s). Keep in mind, setting no end-date effectively means infinite duration! ***Frequency:*** The frequency with which the bookmarked dashboard insights are sent to the recipient(s) (i.e. Daily, Weekly, Monthly, Yearly). ***Time Zone:*** Time zone for which you would like to determine the scheduled time (default UTC). ***Scheduled time:*** The time at which the report, based on the bookmarked insights will be generated and sent to the recipient(s). ***Export format*** The format in which the report is generated and added as an attachment to the e-mail message. ***Recipient(s):*** The recipient(s) that will recieve the bookmarked report in the e-mail inbox. A single or multiple recipients may be selected. Recipients can be selected from the available users within the organisation. Missing a user? follow this guide to add a user to your organisation: [How Can I Add Users to a CM.com Platform Account?](https://knowledgecenter.cm.com/knowledge-center/cm-com-account/user-management/how-can-i-add-users-on-a-cm-com-platform-account.md) ![Screenshot 2025-05-21 at 15.12.57-20250521-131303.png](https://knowledgecenter.cm.com/__attachments/a_89b6b5ede8bc3b33bd4c6efca1a385d58b9b7d0e960ce4ea58de5b2098402a39/Screenshot%202025-05-21%20at%2015.12.57-20250521-131303.png?cb=4fa1e1cf3340add5a53c5a3cac2ee43b) Once you have configured the subscription, you click "Ok" to confirm the subscription. ![Screenshot 2025-05-21 at 10.23.07-20250521-082313.png](https://knowledgecenter.cm.com/__attachments/a_34e3cbd5245399c25505bf50159425cc61470b401c3ab4b0a183a9ca515cc620/Screenshot%202025-05-21%20at%2010.23.07-20250521-082313.png?cb=6b406b872bdc3feea2dd05284422ea78) **Subscriptions management** To see an overview and manage your subscriptions, a new section has been added to the analytics portal. This subscriptions section can be found in the bottom left corner of the screen. ![Screenshot 2025-05-21 at 15.21.37-20250521-132155.png](https://knowledgecenter.cm.com/__attachments/a_037a6cf43f3442b52a7d440e820286fdd6613dff543a6bc44ce78a4824099d07/Screenshot%202025-05-21%20at%2015.21.37-20250521-132155.png?cb=d411f8fbb514b61c01699bceb73f70aa) When navigated to the subscriptions section, you will find an overview of the subscriptions for this account. The overview provides the Bookmark/Subscription Name, Report Name, Recipient(s), From-until dates, Time and Frequency with which the subcription is sent. ![Screenshot 2025-05-21 at 15.26.05.png](https://knowledgecenter.cm.com/__attachments/a_b4691f2c4d66fc261ce1e3ea94c73e8682c6f32cf165364cd67f72d1f414ca31/Screenshot%202025-05-21%20at%2015.26.05.png?cb=439e1a101dfd2be02368f82e07bc18f9) If needed, you can deleted a subscription in order to stop sending it to the recipients by clicking in the delete button behind the three dotted menu item on the right side of the columns. ### **Receiving a report** When a new subscribed report is ready, all configured recipients will receive an e-mail with a link to the bookmarked report and an attachment to directly view the reported view (incl selected filters). ***Example e-mail:*** ![Screenshot 2025-05-21 at 15.32.18-20250521-133223.png](https://knowledgecenter.cm.com/__attachments/a_78a24b7eb848c8caa5c8793b5e06b9e9ee00c38cbaab587d954a030df13cc19f/Screenshot%202025-05-21%20at%2015.32.18-20250521-133223.png?cb=5a5781c31df4743e813bb780ac2c9a14) ***Example attachment:*** ![Screenshot 2025-05-21 at 15.58.31-20250521-135836.png](https://knowledgecenter.cm.com/__attachments/a_62863239577350d6878e9e0a7c378c74f505c0b9dc030ef36ac933859165841a/Screenshot%202025-05-21%20at%2015.58.31-20250521-135836.png?cb=0cf096c3c3dc30eec7c6d9981b6d0eb9) --- language: "en" --- # Conversational Router * [General information Conversational Router](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router/general-information-conversational-router.md) * [FAQ Conversational Router](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router/faq-conversational-router.md) --- language: "en" --- # FAQ Conversational Router * [I cannot login to the Agent Inbox / Mobile Service Cloud...](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router/faq-conversational-router/i-cannot-login-to-the-agent-inbox-mobile-service-2.md) * [Where can I find the router?](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router/faq-conversational-router/where-can-i-find-the-router.md) * [How do I use the router?](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router/faq-conversational-router/how-do-i-use-the-router.md) --- language: "en" --- # How do I use the router? While we are working on adapting the Conversational Router to facilitate people who are less tech-savvy more and more at this point it is still a developer centric product. Therefore, the documentation concerning usage of the router can be found in our [developer documentation](https://developers.cm.com/central-services/docs/about). --- language: "en" --- # I cannot login to the Agent Inbox / Mobile Service Cloud... ## Issue When logging into MSC you get a message you can't log in. The message will look like this: ![Login_failed.png](https://knowledgecenter.cm.com/__attachments/a_2b138f0db6d5282a70388de0d69ef6123825ed2b1b2710a00d38ac178ef94ef3/Login_failed.png?cb=7353b2e4dd666e73703c777a3b163581) ## Product Mobile Service Cloud ## Resolution Please check the entered e-mail address and password and try again. If you are still unable to log in, please select: "Forgot your password?". You will be able to enter the email address you use to log in. You will receive an e-mail to reset your password. ![Schermafbeelding_2022-09-30_154833.png](https://knowledgecenter.cm.com/__attachments/a_03454ce884a878167383958b74038201cbfc266dca4028dab16ad9f401986d61/Schermafbeelding_2022-09-30_154833.png?cb=e0539beab0bf57a12d44aabfc3c8bf0a) ## I don't have access to the inbox of the e-mail address I use to log in: If you do not have access to the inbox of the e-mail address you use to log in, please get in contact with someone within your organization who has access to the settings in MSC. They will be able to change your password. They will be able to change the password by navigating to the Settings ⇾ Team ⇾ \[Your profile\]. Here, you will be able to enter a new password. Make sure you save this new password at the top of the page. ![Password_profile_settings.png](https://knowledgecenter.cm.com/__attachments/a_d3eb4046a89163680afc53ac4edd70c70b38fdd5b9987b0d1975e455eaf9b92b/Password_profile_settings.png?cb=b4f0925f1479b28e4aab85e5c1e640ae) ## My profile has been deleted, can I get it back? When you can not find the profile you are looking for, it may be deleted. Please contact support@robinhq.com. The support representative will be able to reactivate the account. If you have access to the command center, go to user management and set the deleted user's account back to its original account type. (Regular/Collaboration). This will automatically restore the account to it's latest state. ## Background The e-mail address used to log in does not have to be a real/live e-mail address. This can be fictional. We do not recommend to use fictional e-mail addresses for log in. You will not be able to receive the mail to reset your password. --- language: "en" --- # Where can I find the router? The Conversational Router can be found in the nine-tile menu, under the Central Services category. ![Screenshot_2024-11-11_at_17.31.44.png](https://knowledgecenter.cm.com/__attachments/a_4a3e69044056a8542859099f17a08d4f9d015456b47ffe0a7643a67459cd3099/Screenshot_2024-11-11_at_17.31.44.png?cb=02f432986161683ee445fc3cc336c26c) --- language: "en" --- # General information Conversational Router * [Introduction to the Conversational Router](https://knowledgecenter.cm.com/knowledge-center/central-services/conversational-router/general-information-conversational-router/introduction-to-the-conversational-router.md) --- language: "en" --- # Introduction to the Conversational Router The Conversational Router is the message routing platform central to all chat applications within CM.com. It enables bidirectional communication between all connected applications and provides a platform where these applications can benefit from and enhance one another, creating more customer value than would individually be possible. While the Conversational Router is a central piece of CM's infrastructure it is also expanding to integrations with third-parties, giving the benefit of even more additive value to all connected applications. ![Screenshot_2024-11-11_at_17.35.02.png](https://knowledgecenter.cm.com/__attachments/a_73eb3963099e7014a19f7ec06f4248e039dd6c90705dc5a84006c881e309d267/Screenshot_2024-11-11_at_17.35.02.png?cb=96e4d4e80421a351035a790173a9ab1e) --- language: "en" --- # Marketplace These guides provide step-by-step instructions for integrating your existing systems with [CM.com](http://cm.com/)'s platform through our Marketplace integrations. ## Available Guides Click any integration below for detailed setup instructions: * [CRM Integrations](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations.md) * [Live Chat Integrations](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/live-chat-integrations.md) * [Internal Comms Integrations](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/internal-comms-integrations.md) * [CMS Integrations](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/cms-integrations.md) * [Google Ads Integrations](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/google-ads-integrations.md) * [Product Feed Integrations](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/product-feed-integrations.md) * [CDP Integration](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/cdp-integration.md) --- language: "en" --- # CDP Integration --- language: "en" --- # Facebook Ads Integration ## Overview The Facebook Ads integration allows you to sync your AddressBook contact groups to Facebook Custom Audiences. This enables you to target your existing contacts with ads on Facebook. ## Prerequisites Before setting up the integration, ensure you have the following: 1. **Facebook Business Portfolio (Business Manager)** --- You must have an active Business Portfolio. If you don't have one, create it at [business.facebook.com](https://business.facebook.com/). 2. **Facebook Ad Account** --- An Ad Account linked to your Business Portfolio. Note down the Ad Account ID (numeric, found in Business Manager under Ad Accounts). 3. **Admin Access** --- The user authorizing the connection must have admin/full control permissions on the Business Portfolio. 4. **Custom Audience Terms of Service Accepted** --- Before the integration can create Custom Audiences, you must accept Facebook's Custom Audience Terms. * Visit: `>` * This is a one-time requirement per ad account. 5. **AddressBook Groups** --- You must have contact groups created in AddressBook that you want to sync to Facebook Ads. ## Setup Steps ### Step 1: Start Authorization 1. Navigate to the CDP Marketplace and select the **Facebook Ads** adapter. 2. Click **Connect to authorize Facebook Ads**to start the connection process. ![image-20260323-120158.png](https://knowledgecenter.cm.com/__attachments/a_aa354ac645839488ae15e58078a0436e3810e48603cb292e8018bcbaede0fd23/image-20260323-120158.png?cb=2d4c189c84ff15a52361e86ca97676fd) ### Step 2: Select Your Business Portfolio * Facebook will ask: **"Which business portfolio would you like to connect to CDP Ads Integration?"** * Select your Business Portfolio from the list. * Click **Continue**. ![image-20260323-113843.png](https://knowledgecenter.cm.com/__attachments/a_996dc63d085d0a33b8fd3aa642c248611bedffbff40cfb2a3274879550c5df28/image-20260323-113843.png?cb=5bd9cbbf431920e4db0bcd4496b37292) ### Step 3: Grant Ad Account Access * Facebook will ask you to choose which **Ad Accounts** the integration can access. * Select the ad account(s) you want to use for Custom Audiences. * Click **Continue**. ![image-20260323-114102.png](https://knowledgecenter.cm.com/__attachments/a_60a7c20cbe657b1246bfe862d61fb52477a437f68accf087f7394df0cd04ae37/image-20260323-114102.png?cb=b3462ebca83608068ce1bb8e306fd4dd) **NOTE: Please select only the Ads account that you have provided to the delivery team. Do not choose any other account.** ### Step 4: Review and Confirm Permissions * Review the requested permissions (Ad accounts and `ads_management` permission). * Click **Save** to confirm. ![image-20260323-114207.png](https://knowledgecenter.cm.com/__attachments/a_8fde6b535cbb7dfc35c84b4d16f37b4354a8bbe926bc61c31b33a3d812b470f9/image-20260323-114207.png?cb=6607a2d5378b89e8df6ee1981e54c025) ### Step 5: Connection Complete * You will see a confirmation: **"\[Your Business\] has been connected to CDP Ads Integration"**. * Click **Got it** --- you will be redirected back to the Marketplace. ![image-20260323-114249.png](https://knowledgecenter.cm.com/__attachments/a_8fbfecb28714c67671f31655f128ffb9c56b64889c32f2d03ba3993aed4e3225/image-20260323-114249.png?cb=2faeacc3f671ba0a765d32d7cd6791f0) * The integration will generate a System User Access Token that does not expire. You do not need to re-authorize unless you revoke access. ### Step 6: Configure the Adapter After authorization, provide the following configuration: | Field | Description | Example | |--------------------------|---------------------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | Ad Account ID (Required) | Your Facebook Ad Account ID (numeric) | 123456789 | | Sync Schedule (Required) | Cron expression for automatic sync | You can inform the delivery team about when you would like to sync the data, specifying the schedule for example, daily at a specific time, on Sundays, or on the 1st of every month. | | AddressBook Groups | Select the contact groups to sync | Select from the dropdown | **Facebook Ad Account**: An Ad Account linked to the Business Portfolio you intend to authorize. To find your Ad Account ID: * Go to [Facebook Ads Manager](https://adsmanager.facebook.com/) * Click the **Account dropdown** in the top-left corner * Your Ad Account ID is the numeric value displayed below the account name (e.g., `123456789`) * Ensure this ad account is linked to the Business Portfolio you will grant access to during authorization ![image-20260323-121104.png](https://knowledgecenter.cm.com/__attachments/a_c074761ac66f7860ee6546056f57a9d29604f7c29be35c1a602d1b77dbed670c/image-20260323-121104.png?cb=a90937d7fe6edefba989c17ba28cc933) **Manage Groups to Sync** You can click on the **Manage Groups** button, which will redirect you to a page where you can select the groups you want to sync with Facebook Ads. ![image-20260323-121959.png](https://knowledgecenter.cm.com/__attachments/a_a6e5c44c492349d072a7a0454203d449c1cf3adc94bbb96652879690635f635c/image-20260323-121959.png?cb=a0b6d65a67e32fe391d520cc1a4f6f3f) You can select multiple groups that you want to sync and can click on **save groups** button ![image-20260323-122207.png](https://knowledgecenter.cm.com/__attachments/a_d4797fd3aa8b22ae6a52a4d7bb77f8d4b160d83c65126b42ab42bbc416e3df7d/image-20260323-122207.png?cb=00123b406e59fd7172b06d694f768f0f) ### Step 7: Sync Once configured, the integration will: * Automatically sync contacts based on the configured schedule. * You can also trigger a **Sync Data**manually from the adapter settings. ## How the Sync Works 1. **Initialization** --- The system prepares the sync session. 2. **Audience Setup** --- For each selected AddressBook group: * Checks if a Custom Audience with the same name already exists in your Ad Account. * If it exists, reuses the existing Custom Audience. * If not, creates a new Custom Audience with the same name as your AddressBook group. 3. **Contact Sync** --- For each group: * Retrieves contacts from the AddressBook group. * Filters out contacts without email addresses and unsubscribed contacts. * Hashes email addresses using SHA-256 --- raw emails are never shared with Facebook. * Uploads hashed emails to the corresponding Custom Audience in batches of 1,000. ## Important Notes * **Token Validity** --- The integration uses a System User Access Token that **never expires**. You do not need to re-authorize periodically. * **Data Privacy** --- Email addresses are hashed (SHA-256) before being sent to Facebook. Facebook matches these hashes against their own user database --- plain-text emails are never shared. * **Match Rate** --- Facebook matches uploaded hashed emails against its user base. Not all contacts will match, which is expected behavior. Match rates vary depending on how many of your contacts have Facebook accounts with the same email. * **Unsubscribed Contacts** --- Contacts marked as unsubscribed in your AddressBook are automatically excluded from the sync. * **Custom Audience Naming** --- Custom Audiences are created with the same name as your AddressBook groups. * **Re-authorization** --- If you need to change which ad accounts the integration can access, simply go through the authorization flow again. ## Troubleshooting | Issue | Solution | |--------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------| | "Custom Audience Terms not accepted" | Accept the terms at `>` | | "App isn't available" during authorization | Ensure you are logging in with an account that has admin access to the Business Portfolio. If the issue persists, contact support. | | Sync fails with "Access token not found" | Re-authorize by disconnecting and reconnecting the adapter. | | Sync fails with "Unauthorized" | Your token may have been revoked from Facebook Business Manager. Re-authorize the connection. | | "Ad account owner has NOT granted ads_management permission" | Re-authorize the integration and ensure you grant access to the correct ad account. | | Sync completes but no contacts appear | Verify that your AddressBook group has contacts with valid email addresses that are not unsubscribed. | | Low match rate on Facebook | This is normal --- Facebook can only match emails that are registered on their platform. | | Custom audience not appearing in Ads Manager | It can take up to 24 hours for a new audience to become available for targeting. | --- language: "en" --- # Google Ads Integration This guide explains what you, as the customer, need to provide and configure to enable the integration between your Google Ads account with the Marketplace. *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * Admin access to your Google Ads account. * Your **Google Ads Customer ID**. * Access to the Marketplace platform. *** ** * ** *** ## 2. Authorize Google Ads Integration To enable the integration, you need to authorize our app to access your Google Ads account. Follow these steps: ### Steps to Authorize 1. Log in to the **Marketplace** platform. 2. Navigate to the **Integrations \> Google Ads** section. 3. Click on the **Authorize** button. ![image-20251128-084509.png](https://knowledgecenter.cm.com/__attachments/a_65db4c21cac2add5f6805dff53baba6c33e9e336a23e094055d7421fef617dff/image-20251128-084509.png?cb=aa4b8778e52b95f21d992c5ee3d6f014) 4. You will be redirected to the Google authorization page. 5. Log in with your Google account that has access to your Google Ads account. > When you connect your Google Ads account, Google may prompt you to create a passkey. Set one [**g.co/passkeys**](http://g.co/passkeys) up at ahead of time new passkeys need up to 7 days before they work. 4. Grant the necessary permissions to our app. 5. Once authorized, our platform will generate a **Refresh Token** that will be used to access the Google Ads API. ### What to Provide * **Google Ads Customer ID**: This is your unique identifier for your Google Ads account. You can find it in the top-right corner of your Google Ads dashboard. ![image-20251128-084520.png](https://knowledgecenter.cm.com/__attachments/a_506b50bf63eca01f4f1136cef328a5d6a3c7fa94710cceee6c1f2d672aaa4202/image-20251128-084520.png?cb=0191a5a99dac79f3bc47e981b880e745) Copy it without the **Dashes**. *** ** * ** *** ## 4. Manage Groups for Syncing If you want to sync specific groups from your Address Book with Google Ads, follow these steps: ### Steps to Manage Groups 1. In the Marketplace platform, go to the **Integration Details** section under **Google Ads**. 2. Click on **Manage Groups**. ![image-20251128-084539.png](https://knowledgecenter.cm.com/__attachments/a_c99fdbcfae69ef456264b2113dc60df4000858edb7de27c2e6f9dbbf6614ab88/image-20251128-084539.png?cb=74c69425654020e76bbeb7787e077879) 3. Select the groups you want to sync with Google Ads. ![image-20251128-084552.png](https://knowledgecenter.cm.com/__attachments/a_efbcec09022118a06167af1bef39775ff92d977f85bbd2955c2673beaebce68a/image-20251128-084552.png?cb=39e40d77d192e0db9d86ccb9ded94521) --- language: "en" --- # CMS Integrations * [Confluence](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/cms-integrations/confluence.md) * [SharePoint](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/cms-integrations/sharepoint.md) * [Google Drive (Business)](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/cms-integrations/google-drive-business.md) * [CX-One Expert Knowledge](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/cms-integrations/cx-one-expert-knowledge.md) --- language: "en" --- # Confluence This guide explains what you, as the customer, need to provide to enable the integration between your Confluence CMS and our platform. *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * Admin access to your Confluence instance. * Access to the Atlassian account to generate an API Key. *** ** * ** *** ## 2. Retrieve Confluence Base URL ### Steps to Find the Base URL 1. Log in to your Confluence instance. 2. Go to the **Administration** menu and select **General Configuration**. 3. Click on **Edit**. 4. Copy the **Server Base URL** . **Important: For Confluence Cloud, the Base URL must include** `/wiki`**at the end.** * Example: ` *** ** * ** *** ## 3. Generate API Key ### Steps to Generate API Key 1. Log in to your Atlassian account at [Atlassian account](https://id.atlassian.com/manage-profile/security/api-tokens). 2. Click on **Create API Token**. 3. Provide a name for the API token (e.g., "Marketplace Integration"). 4. Select an expiration date for the token (1 to 365 days). 5. Click **Create**. 6. Copy the API token and save it securely, as it will not be displayed again. **Important:** API tokens expire based on the expiration date you set (up to 365 days). When a token expires, the integration will stop syncing and return authentication errors. To avoid disruption, regenerate the token before it expires and update the API Key in the Marketplace platform. **Note:** The user account associated with the email must have **read access** to the Confluence spaces you want to sync. *** ** * ** *** ## 4. Share Details with the Delivery Team Once you have completed the setup, share the following details with the delivery team: * **Base URL** * **API Key** * **Email ID** (used to log in to Confluence) --- language: "en" --- # CX-One Expert Knowledge This guide explains what you, as the customer, need to provide to enable the integration between your CX-One Expert Knowledge CMS and our platform. *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * Admin access to your CX-One Expert Knowledge system. * The following details from your CX-One Expert Knowledge system and share details with the delivery team: * **Base URL** * **Username** * **Key** * **Secret** * **Tag** (to filter the data) *** ** * ** *** *** ** * ** *** --- language: "en" --- # Google Drive (Business) This guide explains what you, as the customer, need to provide to enable the integration between your Google Drive (Business) CMS and our platform. *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * A **Google Business Suite** account. * A **Service Account Email ID** for impersonation. * Admin access to your Google Workspace *** ** * ** *** ## 2. Grant Access to the Service Account Google does not provide an API to approve admin consent, so the admin of your organization must manually grant the Google Drive API access to the Marketplace service account. ### For Existing Customers Only - Remove Previous Domain-Wide Delegation (Optional) If you previously configured domain-wide delegation settings for this integration, you should remove them as they are no longer required. **To remove previous configuration:** 1. Login to [Google Admin Console](https://admin.google.com/) 2. Navigate to **Security** → Search for **API controls** 3. Select **Manage Domain-wide delegation** 4. Locate the marketplace service account **Client ID** and associated **Scopes** 5. **Remove the entry** ![image-20251209-102352.png](https://knowledgecenter.cm.com/__attachments/a_f0bc2df1220bfd45895b85fcbf6ff206143cd801a3825c927c425708107efe41/image-20251209-102352.png?cb=151e12e30b73baa37aa5bc8a77604e32) **Note:** New customers can skip this step entirely. This is only applicable if you have previously set up domain-wide delegation for our service account. ⚠️ **Important: Use "My Drive", not "Shared Drive" (Team Drive)** The files and folders you want to integrate **must live in a personal Google Drive ("My Drive")** and be shared directly with the service account by email. **Items inside a Shared Drive (Team Drive) are not supported** --- even if the service account is added as a member of the Shared Drive, the integration will not be able to read its contents. **How to tell which one you have:** in the Google Drive web UI, files under "My Drive" in the left sidebar are personal; anything listed under "Shared drives" is a Team Drive and will not work with this integration. **If your content is currently in a Shared Drive:** copy or move the file/folder into a team member's "My Drive", then share it directly with `gdrive-accessor@connect-core-prod.iam.gserviceaccount.com` as Viewer. ### Grant Access to Specific Files and Folders (Recommended) Instead of broad domain-wide access, you will now share specific files and folders directly with the service account. **Steps:** 1. Navigate to the Google Drive files and folders you want to integrate 2. Share each file or folder with the **Service Account Email ID** : `gdrive-accessor@connect-core-prod.iam.gserviceaccount.com` 3. Grant **Viewer** access permissions to the service account **Benefits of this approach:** * The service account only has access to explicitly shared files and folders * No unrestricted access to your entire Google Drive *** ** * ** *** ## Next Steps Once you have granted the necessary permissions, notify our delivery team to complete the integration setup. --- language: "en" --- # SharePoint This guide explains what you, as the customer, need to provide to enable the integration between your SharePoint and our platform. *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * Admin access to your SharePoint instance. * Access to your organization's Azure portal to retrieve the Tenant ID. *** ** * ** *** ## 2. Retrieve SharePoint Base URL ### Steps to Find the Base URL 1. Log in to your organization's Office 365 suite. 2. From the nine-tile menu, open the **SharePoint App**. ![image-20251128-070207.png](https://knowledgecenter.cm.com/__attachments/a_a2ed152b406dd2ce5277002707a5995eed17e723dcb5c6d2e9d942181811240f/image-20251128-070207.png?cb=e0ed67865ea467014547c94d390078d7) 3. Copy the **Base URL** of your SharePoint instance. * Example: [https://yourdomain.sharepoint.com.](https://yourdomain.sharepoint.com./) ![image-20251128-070217.png](https://knowledgecenter.cm.com/__attachments/a_230ea04bcb1a95996df9382c06adb018b1ac925b9571dc1c6d04159b34e60053/image-20251128-070217.png?cb=d266b1ef328537d2894b4e835ef84733) *** ** * ** *** ## 3. Retrieve Microsoft Tenant ID ### Steps to Find the Tenant ID 1. Log in to your organization's Azure portal at [Azure Portal.](https://portal.azure.com/) 2. Search for **Tenant Properties** in the search bar. 3. Copy the **Tenant ID** from the Tenant Properties page. ![image-20251128-070237.png](https://knowledgecenter.cm.com/__attachments/a_e5fc2171597623e1e2b0e2a1a32afab643577ab65495261301b31276d61b6f0d/image-20251128-070237.png?cb=3cb97a735375db018081a26d039aaaae) *** ** * ** *** ## 4. Choose Access Mode Our SharePoint integration supports two access modes. Choose the one that best fits your organization's security requirements: | Mode | Permission | Description | Best For | |-------------------------------------------------------------|-----------------------------------|---------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------| | **Sites.ReadAll** (Default) | Broad access | Allows the integration to search and access all SharePoint sites in your tenant. | Organizations that want full access across all sites without additional configuration. | | **ListItems.SelectedOperations.Selected** (Selected Folder) | Restricted access of Folder Level | Allows the integration to access only specific SharePoint Folders or Files that you explicitly grant permission to. | Organizations with strict security policies or compliance requirements. | ### Sites.ReadAll (Default) * No additional setup required after admin consent. * The integration can search and retrieve content from any site in your tenant. **How to use this mode:** * Leave the **Selected Folder** checkbox unchecked in the UI. * Proceed with admin consent. * You will see a consent screen to grant permissions to our application. ![image-20260123-062037.png](https://knowledgecenter.cm.com/__attachments/a_e65abac7a47d76d3f0e94352ed92d92cdd61eb74af8b04b10d5c18c8b5b30ae0/image-20260123-062037.png?cb=2096c3253ee6bb7840cb6108a0151325) ![8d7ca075-cda8-4f6d-bf9d-c462f2007375.png](https://knowledgecenter.cm.com/__attachments/a_bc5c11e9be3c7b492de4b5d9b96fda7c8559da4c3e653ccd6833fc97cfd0e32e/8d7ca075-cda8-4f6d-bf9d-c462f2007375.png?cb=32bd82ab893a90b378176c3cd0a43244) ### Selected Folder (Restricted Access) If you want restricted access at the folder level, select **Selected Folder** by checking the checkbox in the UI before proceeding with admin consent. ![image-20260209-054228.png](https://knowledgecenter.cm.com/__attachments/a_11d3598182f909eb2d1c30816b217ae15648742d4152cd38cf79f7d37b9516c9/image-20260209-054228.png?cb=b410cb161dd19fc0f10cb04ab0dbd80c) **Important:** If you select this option, you will see a different consent screen for folder-level permissions, and you must complete the additional steps below after admin consent. ![image-20260123-062037.png](https://knowledgecenter.cm.com/__attachments/a_e65abac7a47d76d3f0e94352ed92d92cdd61eb74af8b04b10d5c18c8b5b30ae0/image-20260123-062037.png?cb=2096c3253ee6bb7840cb6108a0151325) ![image-20260209-054053.png](https://knowledgecenter.cm.com/__attachments/a_307dc099b262c18ba54d1c69fe9ddea48f64a2647f48e1cdfb4fccf36cf2efb4/image-20260209-054053.png?cb=08209ef5ac0b0370f8b8d4028b42be96) **After that, additional steps are required following admin consent.** #### Step 1: Sign in to Graph Explorer 1. Navigate to [Microsoft Graph Explorer](https://developer.microsoft.com/en-us/graph/graph-explorer). 2. Click on your profile icon at the top of the page. 3. Sign in with your **SharePoint Administrator** account. ![image-20260123-052843.png](https://knowledgecenter.cm.com/__attachments/a_11c11a2829351c8f41d291ad3bb994760bede236b8c49256d32b7d80c16b0392/image-20260123-052843.png?cb=fa9bdedcb1819ea13cc681cbc6d0fa58) #### Step 2: Grant Yourself Full Control Permission To grant site access to our application, you first need to give yourself the necessary permissions: 1. Click on your profile icon and select **Consent to permissions**. 2. In the **Permissions** tab, search for **Sites**. 3. Select **Sites.FullControl.All** and provide consent. **Note:** This permission is only for yourself to enable you to grant access to our application in the next steps. ![78a176a3-ef92-47af-a23e-132e72ce84a2.png](https://knowledgecenter.cm.com/__attachments/a_2c90addf9fa4494013fba01e798aa0690e06e20a40e66a43c48e7a1519d95923/78a176a3-ef92-47af-a23e-132e72ce84a2.png?cb=33d3fde9231e037611851f764b029a0d) ![image-20260123-052946.png](https://knowledgecenter.cm.com/__attachments/a_ed5aa211ca09e583ee9c8556f62f5f76681a1600941e732da62e0b77b3351f9a/image-20260123-052946.png?cb=0c5f22c43bbc1b3046e8d4f3d70aa2f2) #### Step 3: Retrieve the Site ID You need to obtain the Site ID for each SharePoint site you want to grant access to. 1. In Graph Explorer, identify the site you want to grant permission to and note its **Site Name**. **Example:** If your site URL is ` the site name is `Support-Testing-fdsf`. ![image-20260123-053421.png](https://knowledgecenter.cm.com/__attachments/a_53754f58ba1ec71f72d839de6a67b4cf7c5b2cb3fda2fd3b20da97b58e8e4837/image-20260123-053421.png?cb=580cca66eefe9a61dc5eba1d80653572) 2. In Graph Explorer, enter the following GET request in the URL field: https://graph.microsoft.com/v1.0/sites/{hostName}:/sites/{SiteName} **Example:** https://graph.microsoft.com/v1.0/sites/y5x53.sharepoint.com:/sites/Support-Testing-fdsf 3. Click **Run query**. 4. From the response, copy the **id** field value. You will need this Site ID for the next step. ![image-20260123-055842.png](https://knowledgecenter.cm.com/__attachments/a_53f6fb919886bac00b18876cc755d33b2b9ecb1cbace4d089c05d8fb6f388719/image-20260123-055842.png?cb=b3d55f447f9b11de24a80cc98bafd9e5) #### Step 4: Retrieve the Drive ID #### Step 4: Retrieve the Drive ID 1. In Graph Explorer, enter the following **GET** request: https://graph.microsoft.com/v1.0/sites/{site-id}/drives 2. Click **Run query**. 3. From the response, copy the `id` field from the first drive in the `value` array. ![image-20260209-055300.png](https://knowledgecenter.cm.com/__attachments/a_cf615ef2b41f729749d20fbf3dbe5c046119961a629ad9012f5a0c39aaf54a73/image-20260209-055300.png?cb=7419789d7b27ede445b9ecf4268dcd88) *** ** * ** *** #### Step 5: Retrieve the Folder ID Get the folder ID for the specific folder you want to grant access to: 1. In Graph Explorer, enter the following **GET** request: https://graph.microsoft.com/v1.0/drives/{drive-id}/root:/{folder-path} https://graph.microsoft.com/v1.0/drives/b!abc123/root:/MaxCorp 2. Click **Run query**. 3. From the response, copy the `id` field value. This is the Folder ID. ![image-20260209-055507.png](https://knowledgecenter.cm.com/__attachments/a_a1cd237a4dde2f00eaaef633377740f50c0c91bb3061da8a5ac4384d317eb214/image-20260209-055507.png?cb=48b93b85417efdb71c9096bc2d0fa318) *** ** * ** *** #### Step 6: Receive Application Client ID The delivery team will provide you with the **Application (Client) ID** of our Azure AD application. *** ** * ** *** #### Step 7: Grant Folder Access to the Application You must grant our application access to each specific folder using Microsoft Graph Explorer: 1. In Graph Explorer, change the request method to **POST**. 2. Enter the following URL: https://graph.microsoft.com/v1.0/drives/{drive-id}/items/{folder-id}/permissions 3. In the **Request body** section, paste the following JSON: { "roles": ["read"], "grantedToV2": { "application": { "id": "", "displayName": "CM Marketplace SharePoint Integration" } } } 4. Make the following replacements: * Replace `{drive-id}` with the Drive ID you retrieved in Step 4 * Replace `{folder-id}` with the Folder ID you retrieved in Step 5 * Replace `` with the Client ID provided by the delivery team 5. Click **Run query** to execute the request. 6. **Repeat Steps 3-7 for each folder** you want the integration to access. *** ** * ** *** ### Example: Granting Access to Multiple Folders If you want to grant access to folders in multiple sites: | Site | Folder | Access | |-----------|---------------------|-----------| | Max Corp | Personnel info | NO ACCESS | | Max Corp | Financial info | NO ACCESS | | Max Corp | HALO Knowledge base | ACCESS | | Max Works | Other folders | NO ACCESS | | Max Works | HALO Center | ACCESS | For the above scenario, you would: 1. Grant folder permission for "HALO Knowledge base" in Max Corp site 2. Grant folder permission for "HALO Center" in Max Works site **Below is the video for Reference:** [SharepointRestrictedAccess1 (1).mp4](https://knowledgecenter.cm.com/__attachments/a_8c8f1e65229a13ac407b5aed216462942db80630838f0776b78bfd74dcfdae21/SharepointRestrictedAccess1%2520(1).mp4.md) *** ** * ** *** ## Share Details with the Delivery Team Once you have completed the setup, share the following details with the delivery team: | Information | Required | Description | |-----------------|----------|-------------------------------------------------------------------------| | **Base URL** | Yes | Your SharePoint base URL (e.g., `https://yourdomain.sharepoint.com`) | | **Tenant ID** | Yes | Your Azure AD Tenant ID | | **Access Mode** | Yes | Choose: **Sites.ReadAll** (default) or **Selected Folder** (restricted) | --- language: "en" --- # CRM Integrations * [BloomReach CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/bloomreach-crm.md) * [CDP CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/cdp-crm.md) * [HubSpot CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/hubspot-crm.md) * [Magento CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/magento-crm.md) * [MS Dynamics Business Central CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/ms-dynamics-business-central-crm.md) * [SAP CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/sap-crm.md) * [WooCommerce CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/woocommerce-crm.md) * [Returnless CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/returnless-crm.md) * [Shopify CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/shopify-crm.md) * [Zendesk Sales CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/zendesk-sales-crm.md) * [Zendesk Ticketing CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/zendesk-ticketing-crm.md) * [Salesforce CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/salesforce-crm.md) * [TicketMaster CRM](https://knowledgecenter.cm.com/knowledge-center/central-services/marketplace/crm-integrations/ticketmaster-crm.md) --- language: "en" --- # BloomReach CRM This guide explains what you, as the customer, need to provide to enable the integration between your BloomReach CRM with Marketplace *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * Access to your BloomReach Engagement account. * Admin permissions to generate API credentials. *** ** * ** *** ## 2. Provide Your Bloomreach API Credentials To connect your Bloomreach CRM to our platform, you need to generate and share the following API credentials: ### Steps to Generate API Keys 1. Log in to your **Bloomreach Engagement** account. 2. Navigate to **Settings** (gear icon) in the top-right corner. 3. In the left sidebar, click on **"Project Settings"**. 4. Select **"API"** or **"Access Management"**. 5. Click on **"Create New API Key"** or **"Generate API Key"**. ![image-20251128-062352.png](https://knowledgecenter.cm.com/__attachments/a_17e6255aaad7a539c5a71a130b09660a9201b6e88732d38ca2e8aec34c92a320/image-20251128-062352.png?cb=dc6e956e72e9fcfe1bf14b3651bb34e1) 6. Select the required scopes for the integration: * **Customer/Contact Scopes** : `customers.read`, `customers.write` * **Event Scopes** : `events.read`, `events.write`. * **Catalog Scopes** : `catalog.read`, `catalog.write`. * **Campaign Scopes** : `campaigns.read`, `campaigns.write`. * **Consent Scopes** : `consent.read`, `consent.write`. 7. Click **"Generate API Key"**. 8. Copy and securely store the following details: * **API Key ID** * **API Secret** * **Project Token** * **Base URL** > ⚠️ **Important** : The **API Secret** will only be displayed once. Ensure you copy and store it securely. *** ** * ** *** ## 3. Test the Integration Once you've provided the required credentials, our team will test the integration to ensure everything is working correctly. --- language: "en" --- # CDP CRM This guide explains what you, as the customer, need to provide to enable the integration between your CDP CRM and our Marketplace. *** ** * ** *** ## 1. Prerequisites Before proceeding, ensure you have the following: * Access to your CDP account. * Access to the Marketplace platform. *** ** * ** *** ## 2. Generate API Token and Tenant ID To enable the integration, you need to generate an API token and provide your Tenant ID. Follow these steps: ### Steps to Generate API Token and Retrieve Tenant ID 1. Log in to your **CDP** account. 2. In the left navigation menu, click on the **Settings** icon. 3. Under the **General** tab, scroll down to the **API Token** section. ![image-20251128-062651.png](https://knowledgecenter.cm.com/__attachments/a_40549caed4b3574cff2952a70798339d977cee74386b75b99bd274e7e5b88c64/image-20251128-062651.png?cb=444d77e6ee200adc2a0c2697972d8e4e) 4. Locate your **Tenant ID** and copy it. Save it securely for later use. 5. Click on **Create API Token**. 6. In the token creation form: * **Role** : Select **Default**. * **Name** : Enter a descriptive name, such as `CDP Integration with Marketplace`. ![image-20251128-062710.png](https://knowledgecenter.cm.com/__attachments/a_e8bdda72c925b1bbbff057cadd5f2038d4f29d11c9ba6a4f65ac692a6957190e/image-20251128-062710.png?cb=ff0118471cc737ddf01bc21d096314ff) 7. **Generate the token and copy it. Save it securely, as it will not be displayed again.** ### What to Provide * **Tenant ID** * **API Token** *** ** * ** *** ## 3. Retrieve Profile Type ID To enable profile data synchronization, you need to provide the **Profile Type ID**. Follow these steps: ### Steps to Retrieve Profile Type ID 1. In your CDP CRM account, navigate to the **Profile Insights** section from the side navigation menu. 2. Open any profile. 3. In the browser's URL bar, locate the **Profile Type ID** after the `/` in the URL. * Example: If the URL is `profiletypes/12345`, the **Profile Type ID** is `12345`. 4. Copy the **Profile Type ID** and save it securely. ![image-20251128-062750.png](https://knowledgecenter.cm.com/__attachments/a_4cc4746dbbe8e989feaf93f38d39df4b338b8822669100c7e71675495b4ed2f6/image-20251128-062750.png?cb=9f3e9387188350c7d2e02d9ea35c0e8a) *** ** * ** *** ## 4. Share Details with the Delivery Team Once you have the following details, share them with the delivery team: * **Tenant ID** * **API Token (Product Token)** * **Profile Type ID** [Next Page](https://knowledgecenter.cm.com/llms-full.txt/1)