All products

Define product requirements with AI

Clarify what a selected feature should do and record the choices that would otherwise surface halfway through implementation.

You can define product requirements with AI when a feature name masks two different expectations. “Repeat order” could mean buying the same items straight away or reopening the old basket so the buyer can change quantities. The ce-brainstorm skill investigates that difference before a team commits to the build.

The conversation focuses on the person using the feature, the outcome they need and decisions that materially change the work. A visual comparison is available when a layout affects the choice. The result can be passed on as a requirements document with the unresolved questions still visible.

What it does

  • Identify which interpretation of a feature the team intends.
  • Explore what should happen when an earlier item is unavailable.
  • Separate the first release from possible follow-up work.
  • Compare two interactions that lead to different user outcomes.
  • Hand an agreed definition to the person preparing the delivery plan.

Install and use

  1. Use Compound Engineering in a supported agent and supply authorised project context. Use chat or a local preview for private material. A preview exposed on a reachable network port needs a trusted network.
  2. Describe the repeat-order request and name a known disagreement about its behaviour.
  3. Review the choices and approve the scope before handing it to planning.

FAQ

Source

Made by Every. Original skill: ce-brainstorm, published on GitHub under the MIT licence.

View on GitHub