Pilot programme

Ten weeks, one decision, a back-test you can check

We are looking for a small number of design partners for 2027. Each pilot runs on one programme and one decision, and starts by predicting outcomes you already know.

Benefits

What design partners get

Influence

Shape the products

Your questions decide which therapeutic areas and features come first.

Evidence

A back-test on your data

See how well the model predicts outcomes you already have, before relying on it.

Terms

Design-partner terms

Pilot terms reflect the work you put in. Terms are agreed per partner.

The plan

Week by week

WeeksWhat happensYour time
1Agree the decision, endpoints and back-test outcomesTwo workshops
2–3Data agreement, ingestion and cohort buildData team, a few hours
4–5Back-test on known outcomesOne review
6–8Simulate the live decision; design searchOne review per week
9–10Report, evidence package and read-outOne read-out meeting

Fit

Who the pilot suits

  • A live programme with a decision in the next 12 months
  • Phase 1 data for the compound, or outcomes for close analogues
  • A modeller or statistician who can review results
  • Ability to sign a data agreement within four weeks

Nominate

Nominate a programme for a 2027 pilot

Tell us about the decision you want to test. We reply within three working days. The form is not connected yet, so you can also email pilots@sanenza.com.

Not ready to nominate? Request a briefing first.

So we know who to reply to.
Use your company domain.
Sponsor, CRO or academic centre.
For example oncology or rare disease.
For example dose selection for a phase 2b study.
Needed to judge the model fairly.

Questions

About the pilot

What does a pilot cost?

Terms are agreed per partner and reflect the work both sides commit. Ask us for draft terms.

Who owns the results?

You own your data and the report. Sanenza keeps the right to improve its general models only if you agree in writing.

What if the back-test fails?

Then we stop before the live decision and tell you why. You keep the back-test report.