Evidence before engineering

Run the assumptions before you build the product.

Describe your app idea. AssumptionRun identifies its riskiest assumptions, creates a realistic experiment, and shows you what real users actually do.

3 controlled Run types
Behaviour, not hype
Private by default
Human evidence

One controlled loop

From belief to observable behaviour.

AI can make almost any idea sound convincing. That is not evidence. A Run turns one commercial assumption into an observable task with a pre-agreed threshold.

01

Describe your idea

Messy notes are welcome. Confirm the interpretation before anything is tested.

02

Choose an assumption

See the commercial uncertainties most likely to break the idea.

03

Generate an experiment

Create a controlled demand, workflow, or positioning Run.

04

Share with real people

Publish one private link. Testers need no account.

05

Review the evidence

See behaviour, limitations, objections, and a responsible next decision.

Choose how you run

Move alone. Or add a human checkpoint.

Self-Service

Fast evidence, founder-led.

Build and publish a controlled Run without waiting for an expert.

  • Editable assumptions
  • Controlled prototype
  • Automated evidence report
Expert-Assisted

A second pair of sceptical eyes.

The same workflow pauses for human review at the plan, prototype, and report.

  • Run Plan review
  • Test-integrity review
  • Expert-approved conclusion

Build smaller. Learn sooner.

Your next decision should cost less than the product.

Start a New Run