Skip to content

A product manager, after a launch

You shipped a feature and want to know how it landed. The usual route is a form tool, an export and a deck. Here, the answers stay data your agent can read.

A scenario, not a customer story. No company, person or figure here is real; every step is something the product does today.

Where they are

A feature went out last week. You want ten minutes of honest feedback from the people using it, and then you want to know what it says without reading every answer twice.

What they ask for

Copy it and paste it at the agent you already use. It asks for anything it needs to know.

  • A feedback survey for one feature
  • What are users saying?

What the agent builds

A public survey whose answers only you can read, and a launch checklist beside it.

Feedback: the new featureAnyone with the link

What a user answers

How did it go?
What was confusing?
What would you change?

Launch checklist

  • Survey written(done)
  • Link sent to users(done)
  • Dashboard built(to do)
  • Themes read back(to do)

Answers by day, on your dashboard

A schematic, not a screenshot. Your agent writes the page, so no two look alike.

What happens next

  1. Start from a template

    The marketplace has survey templates. Install one and your agent builds from it rather than from a blank page: the structure and the data schema come with it, so the agent writes less and spends fewer tokens getting there.
  2. Share the link

    Anyone with the link can answer without an account. Each answer arrives as a row that only you — and your agent — can read.
  3. Watch the answers arrive

    Put a chart of the answers on a dashboard. Its widgets re-run their query on a short interval, so it fills in while the survey is open.
  4. Ask what it means

    Your agent reads the answers and counts them by theme. The question you would have built a pivot table for is one you ask in a sentence.

What it replaces

  • A form tool, a spreadsheet export and a deck of the results
  • Reading every answer twice to find the pattern
  • A launch checklist in a doc nobody updates

What it does not do

  • It does not email you when an answer arrives. The dashboard and your agent are where you look.
  • Anonymous answers count one per browser, not one per person. A per-person count needs respondents to sign in.

The same move, elsewhere

A roadmap review is the same shape turned around: the agent writes the page, and the team's votes are the form.

Your first artifact is one conversation away

Connect the agent you already use, describe what you want, and it is live at a URL a reviewer opens on a phone.

No MCP support in your agent? Install the skill instead

The other case studies

Product: feedback that stays data · Super Artifacts