# Ask Your Agent Anything, Get It Done

Ask your agent to compare tests, launch a hypothesis, or find a result, and it gets it done. It also reads your data to answer questions and help you decide what to do next.

Published: 2026-07-07

Category: feature

Your agent can now do just about anything you'd open the app to do. Ask it to compare two tests, launch a hypothesis, or track down a result, and it handles the whole task for you, right where you're already talking to it.

It does more than check things off a list. Because it reads your tests, results, and audiences, it can explain what's happening and recommend what to do next, so you get a second opinion backed by your own data every time you ask.

You stay in control throughout. Before it starts, pauses, or stops a test, it checks with you first, so nothing changes without your say-so.

- Compare tests and results in a clean metrics table, right in the chat
- Start, pause, resume, stop, or queue a test from chat, with a confirmation step first
- Create, edit, reorder, and archive hypotheses, or build a test straight from one
- Search your workspace and jump straight to the test, result, or screen you need
- See inline cards for the tests, results, and audiences the agent references, so you always know what it's talking about

[HTML](https://www.pagent.ai/changelog/an-agent-that-runs-your-tests)