Create intermediate Works with: Any AI

Feature Spec for Engineers and Designers

Table output Has example

Communicate a feature clearly to both.

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

Write a feature specification. Feature: [FEATURE]. Problem and users: [PROBLEM_USERS]. Design status: [DESIGN_STATUS]. Technical constraints I know: [TECH_CONSTRAINTS].

Include: summary, motivation, user stories, detailed behaviour with states, content and copy needs, analytics events to track, permissions, edge cases, open questions, out of scope, rollout plan, and acceptance tests. Use tables for states and events. Distinguish requirements (must) from suggestions (could). Add a section titled Decisions Still Needed with owners and due dates. Keep the language plain and consistent.

Add a section called Questions For Review with the questions that engineers, designers and support should each answer before sign-off, and a template for recording their answers. Include a copy deck table with the places where text appears, and a check that error messages tell people what to do next.

Also add a short glossary section to define terms used in the spec, and a table of the dependencies on other teams with the contact for each. Include a list of the metrics that will show whether the feature works after release, and the date of the first review.

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.

Feature: saved filters
States: empty (no saved filters: show prompt), populated, error (save failed: keep filter and retry).
Events to track: filter_saved, filter_applied, filter_deleted.
Permissions: any signed-in user; max 20 per user.
Out of scope: sharing filters.
Decisions still needed: owner for naming rules (product, due Friday); limit 20 or 50 (engineering).
Acceptance tests: save with a name; reject duplicates; apply and remove.
Copy table: "Filter saved" toast; error: "We couldn't save your filter. Try again."
Rollout: 10% of users for one week, then everyone if error rate stays under 1%. Analytics review at day 14.

• requirements