Decision anatomy
Accountable owner
Board, investment committee, Chief Financial Officer, or delegated portfolio owner
Recurrence
Periodic portfolio cycles plus event-driven reallocation
Reversibility
Low for illiquid or committed capital; higher for staged or liquid portfolios.
Cost of error
Very high, with insolvency, strategic lock-in, governance, or opportunity-cost consequences.
Evidence
- cash flows and financial position
- risk and return forecasts
- strategic fit
- scenario and sensitivity analysis
- liquidity, covenant, and regulatory context
Alternatives
- candidate investments
- portfolio weights
- stage-gated funding
- hold cash
- divest or stop
Objectives
- risk-adjusted return
- resilience
- strategic option value
- liquidity
- mission impact
Constraints
- budget
- liquidity
- risk limits
- covenants
- regulation
- governance and conflicts
Uncertainty
- forecast distributions
- correlation
- tail risk
- regime change
- execution risk
How AI may participate
This is design guidance for the archetype, not a universal maturity requirement or a claim about a particular product.
- Assemble investment alternatives, cash-flow ranges, dependencies, constraints, strategic objectives, and downside scenarios.
- Compare allocate, stage, reserve, defer, or reject alternatives under explicit portfolio and liquidity boundaries.
- Expose which recommendation depends on uncertain forecasts, option value, or correlated risk.
Authority and retained human judgment
AI should support analysis and governed proposal. Capital authorization, fiduciary judgment, strategic commitment, and any rights-affecting consequence remain with the accountable committee or executive.
What accountable people still decide
- Which strategic options, resilience benefits, or obligations are not captured by modeled return.
- Whether assumptions and downside tolerance reflect current organizational authority and policy.
- Whether staging or further evidence is preferable to a false binary approve/reject choice.
Outcome observability
Cash movement is immediate; risk-adjusted outcome and attribution may take years.
Evaluation and trace
- Record expected financial and nonfinancial outcomes, ranges, comparator, and horizon before authorization.
- Track staged milestones, realized value, risk, and subsequent allocation changes without claiming clean attribution when horizons are long.
- Compare recommendation, authorized portfolio, funded amount, execution, and observed outcome as separate stages.
Failure and abstention conditions
- Abstain when alternatives, liquidity constraints, dependencies, or material assumptions are missing.
- Do not turn uncertain long-horizon forecasts into precise rankings without sensitivity analysis.
- Escalate any proposal that exceeds authority, policy, concentration, or irreversible-commitment boundaries.
Synthetic walkthrough
From decision context to governed operating design
A synthetic investment committee must allocate $12 million among capacity expansion, cyber remediation, acquisition diligence, staged product investment, and reserve.
This worked example is synthetic. It illustrates Decisagent's decision-first method; it is not advice, a product claim, or evidence that a deployment achieved an outcome.
Choice set
- Allocate now
- Stage the commitment behind milestones
- Hold reserve
- Defer for more evidence
- Reject the proposal
Fit boundary
Signals of a useful fit
- A named accountable committee
- Cash-flow, risk, and value ranges
- Explicit liquidity, covenant, and concentration constraints
- Defined stage gates and dependencies
- A durable Decision Record for later review
Conditional or poor-fit signals
- A one-off opaque deal
- Unknown covenants, conflicts, or mandatory obligations
- Long outcome horizons with weak attribution
- No credible comparator
- The model is asked to substitute for fiduciary judgment
Evidence contract
| Evidence | Time and freshness | Fail-closed behavior |
|---|---|---|
| Current financial position | Decision-time cash, commitments, liquidity, and effective reporting period | Do not propose an amount when liquidity or commitments are unreconciled |
| Ranged forecasts | Scenario version and horizon | Preserve ranges and confidence; prohibit false point precision |
| Downside and scenario analysis | Current stress assumptions and dependencies | Escalate when material downside cannot be bounded |
| Strategic dependencies | Current approved strategy and milestone dependencies | Do not treat unapproved dependencies as confirmed |
| Liquidity and covenants | Current legal and finance interpretation | Prohibit any recommendation that assumes a covenant override |
| Conflicts and mandatory obligations | Current disclosure and governance status | Abstain when conflicts or mandatory uses are unresolved |
Objectives and tradeoffs
Balance expected value, resilience, strategic option value, risk concentration, liquidity, downside protection, mandatory obligations, and timing.
Authority boundary
Use DA-3 analysis and governed proposal only. AI does not authorize, transfer, commit, or move funds. The accountable committee retains fiduciary judgment, conflict resolution, and authorization.
Abstention and escalation
- Liquidity, covenants, conflicts, or mandatory obligations are unknown
- A material assumption cannot be bounded as a range
- The proposal exceeds the approved decision scope
- Outcome attribution would be represented more strongly than the horizon supports
Required human review
- Investment or capital committee
- Finance and treasury owner
- Legal, security, or domain owner for mandatory constraints
- Conflict and fiduciary governance roles
What not to automate
- Covenant override
- Choice involving an undisclosed conflict
- Transfer or commitment of funds
- Fiduciary approval
- Irreversible commitment outside a staged, authorized boundary
Outcome horizon and evaluation
Record authorization and commitment immediately, review staged milestones over quarters, and evaluate realized risk and value over years without overstating clean attribution.
- Predeclare financial and nonfinancial expected ranges
- Compare staged milestones and downside signals before later tranches
- Record differences among recommendation, authorization, funded amount, execution, and outcome
- Preserve long-horizon attribution limits and intervening portfolio changes
Assess your version of this decision
An archetype cannot determine the right authority boundary by itself. Decisagent Fit separately assesses useful AI participation and supportable action authority from your decision context.
Reviewed relevant products
No product relationship has crossed human review for this decision. Candidate seed links are intentionally excluded.