Signed-in context
The guide is available inside the authenticated product. Its request is associated with the signed-in user and the selected workspace rather than an anonymous public chat.
A guide, not an operator
Meet Mia.
Ask better questions.
Mia is the calm layer between a complicated QA workspace and the person trying to understand it. She explains what Matrix QA and your selected workspace support, then tells you where the underlying evidence lives.
Mia · Guide
workspace-aware assistance
What happened in my latest run?
I can help interpret the selected run’s status, findings, and captured evidence. Open the source report before acting on the answer.
Mia does not execute account actions.
01 / the mental model
Think of Mia as a reading light.
Matrix QA has a lot of moving parts: organizations, workspaces, projects, queued runs, browser actions, findings, artifacts, reports, notifications, and settings. Mia is designed to make that surface easier to understand without quietly turning an explanation into an action.
Her answers are contextual, but they are not omniscient. The signed-in request can include the active workspace, and a run detail page can provide a focused run identifier. That context helps Mia explain the record you are looking at; it does not grant her broader access or authority.
“She can help you understand the evidence. You remain the person who decides what to do with it.”
02 / capability map
What Mia can and cannot do.
The boundary is the feature. A useful guide should be specific about both its reach and its restraint.
She can help with
- Explain Matrix QA features, the run lifecycle, and the difference between a browser journey and a direct API check.
- Inspect the authorized workspace history, projects, notifications, and latest runs even when you are on Overview.
- Use a focused run page as deeper evidence context, without requiring a run URL for ordinary workspace questions.
- Explain projects, test capacity, reports, notifications, settings, and the evidence captured around a run.
- Keep a server-backed conversation history scoped to the signed-in user and selected workspace when that history is available.
- Return a degraded or unavailable response instead of pretending that live guidance is available when the guidance service cannot respond.
She does not
- Start, stop, rerun, approve, delete, or modify a run. Mia is a guide, not an account-action controller.
- Change organization data, workspace settings, billing, credentials, permissions, projects, or notifications.
- Execute a browser journey, click through a target site, make a booking, submit a payment, send a message, or perform a destructive action.
- Act as a security auditor, penetration tester, legal adviser, accessibility certification tool, or guarantee that an application is bug-free.
- Provide a secret, token, password, private credential, payment number, or unredacted sensitive value in chat.
- Treat one answer as a substitute for opening the underlying run, report, screenshot, console signal, network event, or timestamp.
03 / under the hood
Why her answers are bounded.
Mia’s behavior is shaped by the same product boundaries that protect the rest of the console.
Scoped history
Conversation history is loaded for the current user and workspace when available. The client keeps the visible history bounded and sanitizes sensitive-looking content before display.
Focused questions
Questions about one workspace or one run are more useful than vague prompts. On a run page, Mia can receive the focused run identifier as context.
Safe by design
Mia explains and points. She does not execute account actions, browser actions, payments, bookings, destructive workflows, or permission changes.
04 / ask her well
Questions that lead somewhere.
How do I run my first test?
Mia can explain the project setup, authorized target URL, journey instructions, test mode, and the role of a dedicated test account. To actually start a run, use the Run Console yourself.
What happened in my latest run?
Mia can look up the latest authorized workspace run, summarize its status and evidence, and point you to the run detail for the source record.
Where are my notifications?
Mia can point you toward the notifications area and explain common notification states. She cannot mark notifications as read, change delivery preferences, or send a notification on your behalf.
Can Mia fix the bug she finds?
No. Mia can help explain a finding and suggest where to inspect next, but she cannot edit application code, create a pull request, change your workspace, or deploy a fix.
Does Mia know everything in my account?
She can inspect the authorized workspace records needed for a question. A focused run page gives her deeper evidence context, and responses should still be checked against the source record in Matrix QA.
05 / next step
Open the console. Bring Mia a real question.
Start with the workspace you already use, ask about a report or run, and then verify her explanation against the source evidence. The best Mia workflow is collaborative: she helps you orient, while you retain control.