> ## Documentation Index > Fetch the complete documentation index at: https://docs.mindset.ai/llms.txt > Use this file to discover all available pages before exploring further. # Analytics Dashboard > See who uses your agents, what they ask, and how well your agents answer, all inside Agent Management Studio. The Analytics Dashboard turns every agent conversation into reporting you can act on. It answers the questions you actually care about: who is using your agents, what they are asking, which agents are performing, and whether people are getting good answers. It lives inside Agent Management Studio (AMS), so you get it without a separate BI tool or a data team. ## Where to find it Analytics shows up in two places in AMS. The **Analytics** section in the main navigation gives you the full picture across every agent in your app, split across tabs for Overview, Agents, Users, and Knowledge. The **Insights** tab in the Agents area shows the same agent reporting alongside your agent configuration, and you can filter it down to a single agent by name. What you see depends on your role. App admins see everything across the app. Account admins see the data for their own accounts. ## Key features Across those tabs, the reporting covers four areas, plus the controls that apply throughout. Who is using your agents and how consistently. Volumes, trends, and per-agent comparisons. Whether users are getting good answers. Whether your knowledge ingested cleanly. ### Adoption See how many people are reaching your agents and how often they come back. You get total users, users interacting each month, and active days per user, alongside a per-user list so you can find your most engaged people. ### Agent activity Understand how much work your agents are doing. You get total messages and threads, each trending weekly and monthly and broken down per agent, plus your most active users and the top themes your agents are being asked about. This is where you compare agents side by side and spot the ones carrying the load. ### Response quality Get a clear read on whether answers are landing. You get upvote and downvote counts over time and, most usefully, the reasons users gave for each downvote. Drill-out tables list the actual responses that were rated, including the free-text feedback attached to them, so you can go straight from a downward trend to the specific answers that caused it. ### Knowledge health Check that your knowledge is in good shape. You get file and Knowledge Context counts, files per context, and processing state, so you can see at a glance whether anything failed to ingest. ### Filtering and privacy Every dashboard supports filtering down to the slice you care about, for example a single survey, source, or user. Filters are interdependent, so narrowing one selection narrows the options offered by the others and keeps them relevant. Some dashboards ask you to pick a required filter, marked with a red asterisk, before they load data, which keeps large datasets fast. Anonymised mode suppresses user names and email addresses across every table and hides any chart flagged as sensitive. Turn it on when you want to show the same dashboard to a wider internal audience without exposing personal data. ### Tailored dashboards and exports Dashboards are configuration-driven, so which metrics a given app or account sees can be tailored without a platform release. For customers who need a specific report format, we build fully branded dashboards. Tables sort and aggregate in place, and every table exports to CSV or Excel, pulling the full dataset rather than just the page on screen. Any dashboard exports to PDF as a formatted report you can send straight to a stakeholder. ## Use cases for your end users * Show your own admins and team leads how their people are engaging, without building reporting yourself. * Read the most common downvote reasons and the free-text feedback to find where an agent is missing, then fix the prompt or add the missing knowledge. * Track active users, messages, and threads over time to prove value to the business. * Confirm your knowledge ingested cleanly by checking processing state before a launch.