> For the complete documentation index, see [llms.txt](https://help.meltingspot.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.meltingspot.io/english/coach-ai/configuration/configuring-your-ai-coach.md). # Configuring your AI Coach The more context you give your AI Coach, the more relevant its responses will be. This configuration is done from **Settings > AI Coach**, across three tabs: **Persona** and **Context** (global settings, for all your playbooks) and **Playbook** (settings specific to a playbook). {% hint style="info" %} AI Coach and playbook settings are reserved to **Spot admins**. {% endhint %} When you turn on **two-way conversations**, you are asked to confirm. Indexing of all your Spot's content then starts. A status shows its progress: **Indexing**, then **On** (or **Failed** if something goes wrong). When you turn it off, the index is deleted. If you turn it back on later, a full indexing runs again. Restrict access to the AI Coach to specific member groups (up to **50 groups**). * **Coach name**: the name displayed to your users. * **Tone of voice**: choose one or more tones among Professional, Friendly, Technical and Concise. **Professional** is selected by default, and at least one tone is required. * **Avoid topics**: topics to avoid. Only **PDF** files are accepted. A scanned PDF (an image, with no selectable text) cannot be read by the AI Coach. Each document's status updates live: indexing, creating the summary, ready, or failed. A document can fail if it has more than 30 pages, contains no extractable text, or is unreadable. {% hint style="warning" %} A failed document still counts toward the limit of 3 documents. Delete it to free up the slot. {% endhint %} Select the playbook to configure via the selector at the top of the tab. The Open playbook button takes you to its management page ([👉 more information here](/english/coach-ai/configuration/managing-your-playbooks.md)). Enable a fallback to a human or external resource in case of uncertainty: 1-1 discussion, channel, forum, or external link ([👉 more information here](/english/coach-ai/configuration/configuring-the-ai-coachs-fallback.md)). These guidelines add to the global guidelines, they don't replace them. If they conflict, the playbook guidelines take precedence. The playbook's topics to avoid add to those set for the whole Spot. The same rules as in the **Context** tab apply: PDF only (no scanned PDFs), status updated live, and a failed document counts toward the limit of 3 until you delete it. The button is only available when: * your **company name** (**Context** tab) and the **coach name** (**Persona** tab) are filled in, * **two-way conversations** are turned on, * the AI Coach is active on your Spot. {% hint style="info" %} The test window is a sandbox. Each time you open it, a new conversation starts. Content cards cannot be opened and feedback is not offered. Your test conversations are not visible to members. {% endhint %} From this window, the **Open preview as a member** link lets you go further: preview the AI Coach exactly as a specific member of your Spot would see it ([👉 more information here](/english/coach-ai/configuration/preview-as-a-member.md)). The more context you give your AI Coach, the more relevant its responses will be. This configuration is done from **Settings > AI Coach**, across three tabs: **Persona** and **Context** (global settings, for all your playbooks) and **Playbooks** (settings specific to a playbook). ### **Persona tab (global)** #### **Proactivity** * **Proactive messages**: enables the AI Coach's ability to spontaneously suggest content. * **Two-way conversations**: enables free conversation with the user. * **Grounded trainer mode**: enabled by default. The AI Coach first drafts a response using the available resources, then shares those resources for deeper study. {% hint style="warning" icon="eyes" %} If **conversational mode** is disabled, you can still edit all the elements on this page, but they won't be used by the Coach. {% endhint %} #### **Conversation** * **Coach name**: the name displayed to your users. * **Tone of voice**: Professional, Friendly, Technical, or Concise. * **Avoid topics**: topics to avoid. {% hint style="info" %} If you answered the MeltingSpot agent's questions when creating your Spot, these fields are already pre-filled. Edit them at any time from this page. {% endhint %} ### **Context tab (global)** #### **Company** Provide: * **Your company name** * **Website URL** * **Company LinkedIn URL** #### **Guidelines** * **Introduction and objectives**: describe your Spot and its objectives. * **Audience**: who are your target users? * **Maturity level**: Beginners, Intermediate, or Experts. * **Use cases**: how do you want the AI Coach to support your users? #### Documents In this section, you can add up to 3 PDF documents to enrich the sources your AI Coach relies on to answer your users. {% hint style="warning" %} Each document is limited to **20 MB** and **30 pages**. {% endhint %} Once your document is uploaded and indexed, you can view a summary of its content by clicking on it. From the list, you can delete a document at any time: it will then no longer be used by the AI Coach. {% hint style="info" %} These guidelines apply by default to all your playbooks. Complete them with specific context from the Playbook tab. {% endhint %} ### **Playbook tab (per playbook)** Select the playbook to configure via the selector at the top of the tab. The Open playbook button takes you to its management page (👉 more information here). #### **Fallback** Enable a fallbakc to a human or external resource in case of uncertainty: 1-1 discussion, channel, forum, or external link (👉 more information here). #### **Guidelines (specific to this playbook)** Complete the global context with details specific to this playbook: * **Introduction and objectives** * **Use cases** * **Avoid topics** #### Documents In this section, you can add up to 3 PDF documents to enrich the context of a particular playbook. {% hint style="warning" %} Each document is limited to **20 MB** and **30 pages**. {% endhint %} Once your document is uploaded and indexed, you can view a summary of its content by clicking on it. From the list, you can delete a document at any time: it will then no longer be used by the AI Coach. {% hint style="info" %} These context documents only apply to conversations that take place within a particular playbook. {% endhint %} ### **Testing your configuration** The **Test AI Coach** button opens a test window reflecting the actual behavior for the selected playbook, fallback included. From this window, the **Open preview as a member** link lets you go further: preview the AI Coach exactly as a specific member of your Spot would see it (👉 more information here). --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://help.meltingspot.io/english/coach-ai/configuration/configuring-your-ai-coach.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.