01 / Provenance
Keep the source with the evidence.
Evidence should retain its origin, date, context, and relevant limitations. A later revision should not silently rewrite what was available when a decision was made.
Method / A basis for trust
Make the reasoning
open to scrutiny.
Pattern is designed around explicit evidence boundaries, visible uncertainty, and human review. A consequential decision should be understandable at the time it is made and examinable later.
Evidence and models should not be confused
What is known.
What is assumed.
What is decided.
These distinctions are the foundation of Pattern’s intended decision workflow.
02 / Evidence boundaries
Keep different kinds of knowledge distinct.
Observed evidence, reported claims, assumptions, model-derived outputs, and human judgment should carry explicit labels. Conflicting evidence and missing information remain visible.
03 / Explicit assumptions
Show what the analysis depends on.
Scenario inputs, constraints, and institutional priorities should be stated. A conclusion should be read alongside the assumptions that could change it.
04 / Uncertainty
Make room for what is unresolved.
Uncertainty belongs in the comparison of options. Scenario outcomes are conditional on their inputs; research-only models require appropriate domain validation before decision use.
05 / Reproducibility
Make the analytical basis reviewable.
Where appropriate, the evidence version, assumptions, and analytical context should be preserved so a reviewer can reproduce or examine the basis for a result.
06 / Human review
Keep accountability with people.
Institutional decision owners and relevant domain specialists should review evidence, assumptions, trade-offs, and limitations. Model-derived outputs inform judgment; approval remains a human responsibility.
07 / Decision records
Preserve why a choice was made.
The intended decision record connects the selected option with evidence, rationale, approvals, alternatives, and relevant dissent, then supports review as implementation and evidence evolve.
The process
From evidence to review
Evidence
Establish what is known, what is disputed, what is missing, and where the evidence came from.
Options
Define the actual actions available to the institution.
Scenarios
Model how different choices may perform under explicit assumptions and uncertainty.
Trade-offs
Compare cost, expected impact, risk, feasibility, timing, and institutional constraints.
Decision
Capture the selected option and the reasoning behind it.
Review
Monitor outcomes and reassess the decision as evidence changes.
Start a conversation
Working on a decision that is difficult to structure, compare, or defend?
Pattern is interested in institutional pilots, domain validation, research collaboration, and decision-workflow partnerships.
Institutional enquiries