> For the complete documentation index, see [llms.txt](https://help.daasity.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.daasity.com/core-concepts/ai-analyst/ask-in-daasity-ai-analyst-in-app.md). # Ask in Daasity (AI-Analyst in app) Ask the AI Analyst a question in plain English inside Daasity and get an answer from your own governed metrics - no manual SQL, no setup. The in-app AI Analyst is a chat assistant built into Daasity. Ask a question in plain language and it answers using your governed Daasity metrics - the same definitions behind your dashboards and Explores. No manual SQL, no report requests, no waiting on an analyst for a quick number. ### Getting started {% hint style="warning" %} **Unsure if you have access to the AI Analyst?** Easy. Look into your nav to find the stars icon (below) in the top-left of your screen. ![](https://410725593-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FamTMWiPne1v1V3L7mbuj%2Fuploads%2FW0vZovV6isdrf5iOtiYu%2Fimage.png?alt=media\&token=dde2bca1-ec8a-4589-b0bb-b36dc4f54f67) If it's there, you have the AI Analyst (including the [MCP feature](https://help.daasity.com/core-concepts/ai-analyst/connect-your-ai-mcp)) enabled. Dig in, and start prompting. Don't see it? That means that it has either not been enabled for your organization or you in particular. Contact your Account Manager or the Daasity Support team to determine which - either way, they can enable it for you quickly. {% endhint %} Open AI Analyst from your Daasity workspace. You'll see **"How can I help you today?"** and a set of example questions. You can click one to run it, or type your own question in the box at the bottom. Example starters include: * *Top 10 products by revenue last 30 days* * *Revenue by channel last 90 days* * *Daily revenue trend this month* * *New vs. returning customers* Press **Enter** to submit, or **Shift+Enter** to add a new line. Ask in everyday language - you don't need to know table or field names. ### What an answer looks like As it works, AI Analyst shows what it's doing - finding the relevant fields, then running the query - so the process is transparent. The answer itself is written in plain English and typically includes: * **A clear summary** with the headline result called out (for example, the top product with its revenue and units), followed by a ranked breakdown. * **A short insight** interpreting the numbers - concentration, trends, or risks worth noticing. * **Assumption and caveat notes** flagged inline. For instance, if you ask for "last 30 days," it may note that it used the closest standard window (such as the latest 4-week period) and call out anything excluded. These notes are there so you can trust exactly what was measured. * **View data.** Click **View data** under any answer to expand the underlying rows behind the response. * **Daasity Verified.** When an answer maps to a specific Explore in your model, it's marked **Daasity Verified** - you can tie that answer back to an Explore you can open and check. Answers without the badge didn't resolve to one specific Explore; the assumption notes and **View data** show you exactly what was measured. Because answers come from your governed semantic layer, the numbers line up with the rest of Daasity. ### Digging deeper Under an answer, open **Dig deeper?** for suggested follow-up questions tailored to what you just asked (for example, "Which retailers are driving this product's sales?"). Click one to run it, or simply type your own follow-up in the box - the assistant keeps the context of your conversation within the thread. ### Working with threads Each conversation is saved as a **thread** in the left panel under **Recents**, so you can return to a previous question or keep building on it. Start a fresh line of questioning any time with **New thread**, and use the search box to find an earlier conversation. ### Tips for good answers * **Name the time frame** ("last 30 days," "this month," "Q2") - it removes ambiguity. * **Name the dimension you care about** ("by channel," "by product," "new vs. returning"). * **Straightforward data pulls are the sweet spot** - a number, a ranking, or a trend comes back reliably. * **Check the assumption notes** on the answer so you know exactly what was measured. * **For analysis or opinions, keep it open-ended and iterate.** Ask follow-ups and sanity-check the result before you rely on it. ### Related tools in the workspace menu The workspace menu (click your workspace name, bottom-left) includes helpers that support your analysis: * **Query History** - revisit past queries. * **Data Dictionary** - browse the models, metrics, and fields available to ask about. * **Domains** - see the subject areas your questions can draw from. * **Refresh Schema** - pull in recent changes to your data model. ### Prefer to use your own AI tool? If you'd rather ask these same questions from Claude, Cursor, Codex CLI, ChatGPT, or another MCP-compatible tool, [connect it to your Daasity data with MCP](/core-concepts/ai-analyst/connect-your-ai-mcp.md).