Know how the market reactsto your decision, before you ship it.
Describe a business decision in plain language. Hundreds of grounded respondents, each tied to a real demographic segment, react to it — and you get a calibrated, signed verdict in minutes.
The problem
TodayYou test code before you ship it.Your pricing decisions ship on a slide deck.
A price change can cost more than any outage, yet it goes out on a deck and a room's confidence. The usual ways to check arrive too late.
Market research
6–10 weeks
Five figures per study, and it lands after the decision is committed.
Survey panels
2–4 weeks
Stated preference, not reaction. People under-report churn intent.
Limited rollouts
4–8 weeks
A real signal — but the backlash is already public, and attributed to you.
Preditance returns a verdict in minutes, before anything ships.
One paragraph in.One signed verdict out.
Not a chatbot giving an opinion. Hundreds of differently grounded people react, and their disagreement is measured — because the disagreement is the answer.
- Step 1 of 4: Describe the decision. Paste it in plain language — one paragraph is enough. Unusable documents are rejected first, and company names are blinded so agents judge the decision, not the logo. Illustration: the decision drawn as a page of text, with the company name blinded.
- Step 2 of 4: Ground the population. Personas are drawn from real demographic segments by region, income, age and sector. Regulators, journalists and activists are included on purpose. Illustration: the personas gather into nine demographic segments, with a few edge cases between them.
- Step 3 of 4: Let them react. Every persona reacts in character and returns a structured objection, so you can see exactly where resistance concentrates. Illustration: each persona accepts or rejects; rejection is concentrated in one segment.
- Step 4 of 4: Read the verdict. Reactions become a rejection rate with a confidence band and one of five verdicts. The result is signed, so nobody can edit it later. Illustration: the personas line up as a bar, 42% rejecting — cleared with modifications.
- Illustration: the decision drawn as a page of text, with the company name blinded.
Step 1 of 4
Describe the decision
Paste it in plain language — one paragraph is enough. Unusable documents are rejected first, and company names are blinded so agents judge the decision, not the logo.
- Illustration: the personas gather into nine demographic segments, with a few edge cases between them.
Step 2 of 4
Ground the population
Personas are drawn from real demographic segments by region, income, age and sector. Regulators, journalists and activists are included on purpose.
- Illustration: each persona accepts or rejects; rejection is concentrated in one segment.
Step 3 of 4
Let them react
Every persona reacts in character and returns a structured objection, so you can see exactly where resistance concentrates.
- Illustration: the personas line up as a bar, 42% rejecting — cleared with modifications.
Step 4 of 4
Read the verdict
Reactions become a rejection rate with a confidence band and one of five verdicts. The result is signed, so nobody can edit it later.
Illustrative simulation
The output
One reportA verdict, not a report.And the reasons behind it.
Every clearance returns the same anatomy: the call, the number behind it, who resists, what to change — and evidence you can file.
Illustrative simulation of one clearance report. The decision is subscription price increase, tested against 200 agents. A resistance bar is shown for each of four segments — Price-sensitive 77%, Long tenure 46%, Heavy users 22%, Enterprise 14% — so resistance can be seen concentrating in particular segments rather than spreading evenly. The headline is 42% rejection, verdict cleared with modifications.
The call
A verdict and a rejection rate, with how stable that number was across the run.
Who resists
The segments where resistance concentrates, and what they object to.
What to change
A decision memo and three safer variants that keep your strategic intent.
Evidence you can file
A signed certificate with a public link anyone can verify, no account needed.

The record
Counts, not a rateThe record is public,and it is checkable.
Every forward prediction is hash-sealed with its judgement date and published with its resolution criterion. Nobody — including us — can revise it. Hits and misses are both published.
- Sealed publicly
- 3
- Voided as unjudgeable
- 0
- Correct
- 1
- Wrong
- 0
- Still open
- 2
None are past their judgement date without an outcome recorded.
We have audited every prediction on this page, voided the ones we had written badly and corrected the verdicts we had got wrong. Those dispositions are still being applied to the public record, so the statuses below are the ones originally published until they land. The audit sets out what changed and why.
Each one is sealed before its outcome exists, so the list cannot be cherry-picked afterwards. We audited all of them, voided the ones we had written badly and corrected the verdicts we had got wrong — the voids stay published with their reasons. Too few have been judged to quote an accuracy rate, so we publish counts and no percentage.
From the public record, 16 November 2026
Use cases
FiveBuilt for decisionsthe public can reject.
Most teams start with pricing. The same engine clears launches, policy updates, competitive moves and internal change.
Bring your hardest decision.
We'll tell you what the market does next — and sign our name to it.