Observed
Direct evidence from the company, its operators, or a documented deployment.
Can support a decision when source quality and context are clear.
Evidence and decision methodology
Evidence is useful only when you can inspect the reasoning.
Playbook Atlas does not turn case studies into promises. We classify what is known, expose what is assumed, and design the pilot around the unknowns that matter most.
01Evidence classes
Four labels prevent confidence from outrunning the source.
Transferable
Evidence from a comparable setting with the differences made explicit.
Supports a hypothesis, not a promise of identical results.
Assumption
A working belief required to estimate value, feasibility, or adoption.
Must have an owner and a way to test it during the pilot.
Unknown
A decision-relevant gap that the available evidence cannot resolve.
Stays visible in the recommendation instead of being written away.
02Decision process
The evidence trail follows the decision from question to pilot.
Frame
Define the decision, sponsor, operating constraint, economic boundary, and deadline.
Collect
Bring together company facts, operator input, existing systems, and external deployment evidence.
Classify
Label every material claim as observed, transferable, assumed, or unknown.
Compare
Score opportunities on the same dimensions and record disagreements instead of averaging them away.
Design
Turn the leading option into a workflow, control model, measurement plan, and smallest credible pilot.
Red-team
Challenge the value case, evidence chain, failure modes, and reasons not to proceed.
03Opportunity score
A score opens a conversation. It does not replace one.
01
Value
The size and strategic importance of the constrained outcome.
02
Feasibility
Data, integration, workflow, and capability readiness.
03
Evidence
The quality and relevance of support for the mechanism and outcome.
04
Adoption
The behavior change, ownership, and operating model required.
05
Risk
Customer, legal, security, safety, and reputational exposure.
06
Learning speed
How quickly a bounded pilot can resolve the critical unknowns.
04What the method refuses
Clarity also means knowing what not to claim.
No borrowed certainty
A result in another company is not presented as your forecast.
No invisible assumptions
Material estimates remain attached to their assumptions and owners.
No automation theater
A pilot must name the workflow, human handoffs, control points, and success measure.
Apply the method to a real decision
Bring the operating question. Leave with a decision record.
The 30-minute fit review tests whether the question is bounded enough—and important enough—for the sprint.
Request a fit review