Proposed ninety-day pilot
Measure the decision value before anyone scales the promise
The pilot is a paid, scoped engagement with an accelerator, corporate venture group, venture studio, strategic, or development consultancy. It is designed to test whether Decidian produces better-supported decisions with less expert effort and no increase in false confidence.
Phases
Ninety days, three phases
Days 0-15
Legal, evidence, authority, data, and baseline
Agreements, data-rights and consent terms, reviewer authority and conflict records, portfolio scope, and the measurement baseline are established before any analysis begins.
Outputs
- Executed data-processing and confidentiality terms
- Named reviewers with verified authority and recorded conflicts
- Defined portfolio subset and evidence-access scope
- Baseline of cycle time, expert hours, rework, deficiencies, forecast error, and capital spent before blockers surfaced
Days 16-35
Historical replay
Decidian is run against projects whose outcomes the customer already knows, so conclusions can be compared with what actually happened.
Outputs
- Decision objects produced for completed historical projects
- Comparison of surfaced blockers against blockers that materialized
- Initial view of false-blocker and missed-blocker behaviour
Days 36-60
Blinded challenge
Reviewers assess projects without seeing platform output, and the two are compared afterwards, so the platform cannot anchor expert judgment.
Outputs
- Independent expert conclusions recorded before disclosure
- Agreement, disagreement, and unique-finding analysis
- Evidence-traceability audit on every material conclusion
Days 61-90
Live shadow mode
Decidian runs alongside live decisions without authority over them, and the recorded gate results are compared with the decisions the organization actually takes.
Outputs
- Live decision support with no binding authority
- Expert-hours-per-decision measurement against baseline
- Calibration of forecast ranges against interim outcomes
Every measure named in this programme is a pilot measurement target, not an achieved result. Conversion to an ongoing engagement happens only after the agreed evidence and performance gates below are satisfied.
Hypotheses
What the pilot is designed to test
- Blockers are surfaced earlier than in the baseline workflow.
- Expert hours per decision fall without a fall in decision quality.
- Forecast ranges are better calibrated against observed outcomes.
- False-blocker rate stays within an agreed tolerance.
- Every material conclusion in the pilot output is evidence-linked.
- Reviewers report that abstention and contradiction handling increase, rather than reduce, their trust.
Boundaries
Boundaries of the first release
- Every KPI in this programme is a pilot measurement target, not an achieved result.
- Conversion to an ongoing engagement occurs only after the agreed evidence and performance gates are satisfied.
- Pricing is scoped per engagement; figures appearing in planning material are hypotheses requiring validation.
- Pilot data remains the customer's. Cross-project use requires separate written authorization, appropriate de-identification, and contractual authority.
Conversion gates
What must be true before the pilot becomes an engagement
Conversion is gated on measured results, agreed in advance with the customer, rather than on satisfaction with the demonstration.
- Agreed evidence-traceability threshold met on sampled conclusions.
- Blocker-detection performance at or above the agreed threshold, with false blockers within tolerance.
- Measured reduction in expert hours per decision, with no measured loss of decision quality.
- Reviewer-acceptance criteria satisfied by the named expert panel.
- Security and data-handling review closed without an outstanding material finding.
Apply for a pilot cohort place
Cohorts are limited so that every decision produced can be reviewed in depth.