> For the complete documentation index, see [llms.txt](https://docs.screendesk.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.screendesk.io/debug-with-ai/ai-resolution.md). # AI resolution A customer's recording on a ticket or conversation becomes a private note with a suggested fix for your agent. When a customer's recording lands on a ticket or conversation, AI resolution looks into it and leaves your agent a private note: what went wrong and a suggested fix. Customers never see it, and it never replies or closes the ticket for you. {% embed url="" %} Available on Pro and Enterprise, for Zendesk, Intercom, Help Scout, HubSpot, Freshdesk, Freshchat, Jira Service Management and Front. *** ### What it needs Open **Settings → AI resolution**. The **Settings** tab lists what's missing:
Settings, AI resolution, Settings tab
| Need | How to get it | | ----------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Helpdesk** | Connect one so AI can read tickets and add private notes, for example [Zendesk](/helpdesk-integrations/zendesk/install-and-set-up.md) or [Intercom](/helpdesk-integrations/intercom/install-and-set-up.md). | | **Recording summaries** | Turn on AI processing and report intelligence in [Settings → AI](/debug-with-ai/overview.md). | | **Instructions** | Publish them under **Instructions** (below). | | **Knowledge** | At least one trained help center or doc in [Knowledge](/debug-with-ai/data-sources.md). | | **Solved tickets** | Optional: past tickets as extra context. | Once everything is ready, turn on **Investigate tickets automatically**. In Zendesk it works alongside **Tell customers their recording arrived**: that reply is public, its note stays private. In Intercom the note is an internal note in the conversation, and Screendesk admins also see the result in the [inbox sidebar](/helpdesk-integrations/intercom/messenger-editor-recommended.md). In Front the note is an internal comment, and Screendesk admins also see the result in the [Screendesk app](/helpdesk-integrations/front/in-conversations.md). ### Instructions Under **Instructions**, click **Edit** and fill in what applies: * **How it writes**: tone and detail. * **What it checks first**: what it must see before suggesting a fix, such as the app version. * **What it never claims**: statements and actions to stay away from. * **When a person takes over**: cases it hands to your team. * **What it leaves out**: personal or internal details it never repeats. Publishing creates a new version. Earlier versions stay in the history. ### How it decides * **Confidence needed** (75% recommended): below it, the note asks your team to look instead of suggesting a fix. * **Tag tickets in Zendesk**: optional tags so you can track outcomes (Zendesk only). * **Wait for the recording** and **Wait for the transcript**: how long it waits for the summary and what was said before going ahead. *** ### History The **History** tab lists every ticket it looked into, with its outcome and the note it left. Filter by outcome or date and open one to see why. | Outcome | Meaning | | ------------------------ | ------------------------------------------------------------------------------------------------------------- | | **Waiting for evidence** | The recording's summary isn't ready yet | | **Investigating** | It's reading the ticket and evidence | | **Recommendation ready** | It found a fix backed by your docs | | **Needs human** | Your team should look | | **Note posted** | The private note is on the ticket | | **Superseded** | The ticket changed, so a newer attempt took over. After two, it posts what it found instead of starting again | | **Failed** | Something stopped it. Open it to see why. | {% hint style="info" %} No note on the ticket? Check **What it needs** first, then the ticket's attempt in **History**. {% endhint %} --- # 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://docs.screendesk.io/debug-with-ai/ai-resolution.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.