ASYNCHRONOUS RISK ASSESSMENT
Map one risky policy—without a sales call.
Answer the questions below in writing. DryRun prepares a proposed pilot scope for human review. No meeting, integration, or customer data is required.
If the scope fits, the 30-day founding pilot costs €1,500 and includes one policy, one representative anonymised dataset, one controlled simulation, and a decision-ready `DEPLOY`, `REVISE`, or `BLOCK` report.
EXACTLY WHAT YOU RECEIVE
Six concrete pilot deliverables.
No vague transformation project. Every item is written, reviewable, and tied to the one policy you bring. At handoff, you receive the written scope and intake receipt, a secure report view, and a downloadable snapshot-bound JSON report.
- 01 · Pilot charterA written record of your policy, decision owner, safety boundary, acceptance criteria, required fields, and agreed dataset scope. Use it to verify what will—and will not—be tested before data intake.
- 02 · Intake evidenceA receipt showing rows received, imported, duplicated, rejected, and missing evidence, including field-by-field coverage. Use it to judge whether the dataset can support the requested claims before any recommendation is calculated.
- 03 · Reproducible simulation recordA dataset SHA-256, case count, policy version, evaluator version, flagged-case count, and every per-case decision. Use it to identify the exact evidence and logic behind the result and compare a later rerun.
- 04 · Decision reportOne explicit `DEPLOY`, `REVISE`, or `BLOCK` recommendation with false-automation rate, policy violations, risk indicators, estimated—not realised—value range, and limitations. Use it to make a documented policy decision without treating a projection as an outcome.
- 05 · Safer next-step configurationShown inside the completed secure report and included in its JSON export: recommended confidence and approval thresholds, future shadow limit, mandatory human-review boundaries, stop conditions, risks, limitations, and estimated value. Use it as the controlled change list; it is not a production deployment.
- 06 · Client acceptance receiptAn immutable receipt, bound to the exact delivered simulation, that you control after reviewing the report. It proves which report was received and reviewed; it does not approve production deployment, automation, or financial action.
Before anything runsYou explicitly confirm that the imported cases represent the agreed scope and that the displayed policy is the exact rule to test. DryRun records both SHA-256 hashes, the case count, your identity, and the timestamp; any later change blocks execution until you accept a new snapshot.
Your decision at the endYou leave with evidence to choose one of three paths: proceed only within the documented guardrails, revise and rerun, or stop the policy. A `BLOCK` result is a valid paid outcome when the evidence shows the proposed policy is unsafe.
See the output before you buyOpen the interactive synthetic report preview → It uses sample data in your browser and shows the decision, metrics, flagged cases, segment breakdown, and safety boundary. It is illustrative—not customer evidence or a promise of your result.
Where the handoff appearsYour authenticated workspace shows the accepted scope, run status, recommendation, flagged cases, decision evidence, safer configuration, stop conditions, and limitations. The completed report can also be downloaded as a snapshot-bound JSON record; the optional acceptance receipt remains in the workspace.
Not included in the €1,500 pilotNo production integration, live automation, customer contact, guaranteed savings, or deployment performed by DryRun.
30-DAY DELIVERY
A clear written handoff at every step.
- ScopeWritten qualification, NDA/DPA, acceptance criteria, and dataset schema.
- IntakeOne anonymised export, manually reviewed before activation.
- SimulationObserved deterministic replay with exceptions and evidence reviewed by Peter.
- HandoffSnapshot-bound report with a `DEPLOY`, `REVISE`, or `BLOCK` recommendation by day 30, followed by your optional acceptance receipt.
Questions can be handled asynchronously throughout. Peter can provide a short high-level explanation if needed.