Structure the decision before you optimize the output.
A useful plan, model, brief, or recommendation begins with a clearer decision. DECISION OS separates the decision from the desired deliverable and makes the evidence, assumptions, unknowns, options, conditions, and next action explicit.
What exactly must be decided?
Before producing a deliverable, define the decision in a form a human owner can assess:
- Decision: What choice or commitment is required?
- Decision owner: Who has authority to make it?
- Options: What realistic paths are being compared?
- Constraints: What limits the choice?
- Decision point: When does the decision need to be made?
- Evidence scope: What information is relevant to this decision?
A requested output — a plan, model, deck, memo, or analysis — is not automatically the decision itself.
The context changes. The discipline stays decision-centered.
Market entry — What conditions would justify entry, further validation, or a decision not to proceed?
Investment — Which assumptions, evidence, downside conditions, and decision rights matter before commitment?
Fundraising — What should be validated about need, timing, use of funds, and alternatives before deciding to raise?
Partnership — Which objectives, roles, dependencies, commitments, and exit conditions must be explicit?
Pricing — Which customer, economic, positioning, and implementation assumptions need evidence before a change?
New business — What evidence is enough for a bounded pilot, what remains unknown, and what would make the team wait, pivot, or stop?
Fictional/illustrative contexts only. These examples do not imply sector-specific advisory services or domain authority.
FRAME → DECIDE → EXECUTE
FRAME — define the decision and structure the uncertainty
- What needs to be decided?
- What is known?
- What is assumed?
- What remains uncertain?
- Which options and constraints matter?
DECIDE — make evidence and conditions explicit
- Which evidence could change the decision?
- What should be validated before commitment?
- Which conditions are met, unresolved, or contradicted?
- What can the decision owner responsibly decide now?
EXECUTE — make the next step explicit
- What happens next?
- Who owns it?
- By when?
- What evidence or result should return to the next review?
The method is not a rigid universal workflow. In some cases, the next action is execution. In others, the correct next action is to gather evidence, narrow scope, wait, pivot, or stop.
Keep facts, assumptions, and unknowns separate.
FACT / KNOWN
A statement supported by an identified source or direct observation within a defined scope and as-of point.
Useful context can include:
- source
- as-of date or observation point
- scope
- limitations
ASSUMPTION
A condition being treated as true for the purpose of reasoning, but not yet established as a fact.
An assumption should remain visible until it is validated, rejected, or made irrelevant by the decision.
UNKNOWN
Something not yet established that could affect the conclusion, option, timing, or risk.
Not every unknown requires research. Prioritize the unknowns that are decision-sensitive.
Principle: plausible model output is not automatically verified evidence.
Organize information around the decision.
The following patterns are optional and combinable. They are not products, mandatory templates, or one-to-one equivalents of FRAME, DECIDE, and EXECUTE.
- Evidence & Assumption Ledger — trace statements, status, basis, uncertainty, and follow-up.
- Decision Readiness Gates — show which evidence or conditions still matter before commitment.
- Decision Brief — give the human decision owner a compact view of the decision, options, rationale, uncertainty, conditions, and next step.
Evidence & Assumption Ledger anatomy
Suggested fields
- Statement
- Classification: Fact / Assumption / Unknown
- Source or basis
- As-of / scope
- Confidence or uncertainty note
- Follow-up needed
- Owner, when relevant
The ledger makes provenance and uncertainty visible. It does not certify that evidence is complete or automatically verified.
Decision Readiness Gates anatomy
Suggested fields
- Decision condition
- Evidence required
- Current status
- Remaining gap
- Decision implication
A gate can be met, partially met, unresolved, or not applicable depending on the case. The labels are descriptive, not an automated approval system or universal readiness score.
Decision Brief anatomy
Suggested structure
- Decision
- Decision owner
- Options
- Current evidence
- Key assumptions
- Critical unknowns
- Readiness conditions
- Conditional rationale
- Next action
- Owner / review point
A recommendation should remain conditional when critical evidence is missing.
One coherent fictional example
Fictional example — for illustration only The information below is invented to demonstrate structure. It is not a real client case, a recommendation, a performance claim, or evidence of outcome.
Decision context
Decision: Should a small, bounded pilot for a new business idea begin now, wait for one more validation step, or stop? Decision owner: Business Lead (fictional role) Fictional decision deadline: 30 November 2026 Options: Begin bounded pilot / Run one more validation step / Stop
Ledger excerpt
| Statement | Classification | Fictional basis / as-of | Current status | Follow-up |
|---|---|---|---|---|
| A draft pilot scope exists and can be reviewed by the decision owner. | Fact within the fictional case | Fictional internal scope note, as of 20 Nov 2026 | Known | Confirm final boundaries before commitment |
| One small team can run the pilot without adding full-time capacity. | Assumption | Fictional planning assumption, as of 20 Nov 2026 | Unvalidated | Confirm capacity with the proposed owner |
| Intended users will use the proposed workflow repeatedly. | Unknown | No fictional usage evidence yet | Unresolved | Run targeted user-need validation |
Readiness Gates excerpt
| Condition | Evidence needed | Fictional status | Gap / implication |
|---|---|---|---|
| Pilot scope is bounded and reversible | Agreed scope and stop conditions | Partially met | Final stop conditions still need confirmation |
| Delivery capacity is available | Owner/capacity confirmation | Unresolved | Capacity assumption not yet verified |
| User need is sufficient for a pilot | Targeted validation evidence | Unresolved | Critical decision-sensitive gap remains |
Decision Brief excerpt
What can be decided now? The critical user-need condition and delivery-capacity assumption are unresolved. An unconditional Go is not supported by the fictional evidence.
Current disposition: WAIT for one targeted validation step before committing to the pilot.
Next action: Gather evidence on the unresolved user-need question with a small defined group. Owner: Validation Lead (fictional role) Review point: Reassess this decision after the validation evidence is available.
A five-year plan is not always the first deliverable.
In one business-planning context, the following sequence can be useful:
ASSUMPTIONS → VALIDATION → ECONOMICS → BUSINESS MODEL → 5-YEAR PLAN
The point is not that every decision must follow this order. The point is to avoid adding precision to a model before decision-sensitive assumptions have been made visible and, where necessary, tested.
A five-year plan can be valuable. Its usefulness depends in part on the quality and relevance of the assumptions and evidence that feed it.
Make the next action decision-relevant.
A next action should answer four simple questions:
- Action: What specifically happens next?
- Owner: Who is responsible for it?
- Due / decision point: When is it needed?
- Review: What result or evidence should return to the next decision?
The next action does not have to be execution. It can be further validation, clarification of authority, narrowing scope, or a deliberate wait/stop decision.
Limits and onward path
AI can help organize information, surface inconsistencies, identify assumptions, compare options, and draft structured material. It does not own the purpose, decision criteria, risk tolerance, authority, or final decision.
The method does not guarantee decision quality, project success, speed, compliance, or business outcomes. Its purpose is to make the structure, evidence, uncertainty, and next decision more explicit.
