All products

Write a product hypothesis with AI

Put a proposed change, its audience and a measurable prediction in one document. Agree what would count as evidence before the team starts building.

To write a product hypothesis with AI, start with a decision your team has yet to make. The define-hypothesis guide helps a group describe its prediction before anyone picks a winning design. It asks for a specific change, the people it might affect and a way to tell whether their behaviour changed.

For example, a software team might hope that a shorter setup process helps new customers finish their first task. A rise in completed sign-ups would tell only part of that story. A measure taken after setup may be closer to the outcome the team actually wants.

What it does

  • Turn an idea into a prediction about an observable action.
  • Narrow the audience to a group the proposal genuinely affects.
  • Distinguish the primary success measure from supporting indicators.
  • Choose a guardrail that would reveal an unwanted side effect.
  • Record assumptions that could change how the team interprets its result.

Install and use

  1. Install the skill in your agent and bring a short problem statement plus any existing evidence. The planning step needs no analytics account.
  2. Ask, “We think a shorter setup will help first-time customers finish their first task. What would we need to see to believe it?”
  3. Review the proposed metrics, timing and validation approach with the people who will use the document.

FAQ

Source

Made by Jonathan Prisant (Product on Purpose). Original skill: define-hypothesis, published on GitHub under the Apache-2.0 licence.

View on GitHub