Skip to content

Use case

Turn a discovery checklist into a questionnaire that draws the result.

Publish a guided questionnaire that turns answers into a visual result. Fulcrum’s process map is one, and it redraws with every answer.

Who it’s for: Consultants, agencies, operations specialists

A conversation people can take with them

Discovery usually ends with notes only the consultant can read. A diagnostic page puts the same questions to the client. As they answer, they see their processes and problems take shape, along with what to look at first.

Fulcrum built its process map this way. You work through five steps with a fictional sample company, and the map on the right redraws from the answers. At the end you can download a brief.

Fulcrum published it with artifact.guru on its own domain, where you can open the Fulcrum process map.

How it’s put together

  1. Fixed questions and rules

    The questions, and how each answer moves the result, are written down in advance. The page applies those rules; it doesn’t call an AI model.

  2. Everything stays in the browser

    Answers are kept in memory while the page is open. Nothing is uploaded, and the brief is generated on the visitor’s own device.

  3. One file, one link

    The whole diagnostic is an HTML page. It’s published to artifact.guru and served from Fulcrum’s own domain.

  4. Improved at the same link

    New questions or clearer wording go out as new versions. Anyone who bookmarked it gets the latest.

Limits

  • There’s no database, so answers aren’t collected anywhere. If you need submissions, connect a form service you trust and say so on the page.
  • Results come from the rules you wrote. Call it a starting point for a conversation, not a diagnosis.
  • A complex page takes real design and testing. Expect several rounds with your agent, each one a new version.

Prompt

Make your own

Have these ready, then paste the prompt into your agent:

  • The question list
  • Written rules for how answers change the result
  • Sample answers, labeled fictional
  • The output you want
Prompt
Build a self-contained HTML diagnostic from the questions and rules I attach.
Label the sample answers as fictional. Show how each answer changes the result,
and add Reset and a downloadable summary.

Use only the rules given. Don’t imply an AI reads documents or that results are
validated advice. Keep answers in memory. Don’t upload files, fetch links,
or pretend to submit contact details.

Make it work with a keyboard and at 390px and 1440px wide, with a readable
list as an alternative to any diagram. Test the sample through to the summary.
Publish it privately to artifact.guru and give me the link and what you tested.

Common questions

Can it email me the answers?

Not by itself. Pages run in a sandbox without a server. Link to a form you control if you need submissions, and tell visitors what it collects.

Can it be on our own domain?

Yes, on Individual and Team plans. The Fulcrum example is served from artifact.fulcrum.rocks.

Read next

More use cases

Try it with your own file

Copy the prompt above, add your file and publish.