Before funding
AI portfolio arbitration
Compare options pursuing the same purpose, remove those that are not admissible and show which assumptions actually govern allocation.
Before committing resources
- 01
Frame
One shared objective
- 02
Filter
Ineligible options
- 03
Decide
A recorded choice
Criteria come before ranking.
Engagement at a glance
- Decision
- Select, investigate, defer or remove — with rationale and authority
- Precondition
- One purpose, horizon and baseline shared by all options
- For whom
- Executive leadership, transformation, CIO, business sponsors
- Result
- Rank, uncertainty, switching values and decision retained by version
The question
Which use cases actually deserve time, data and budget?
When several AI ideas compete for the same resources, a weighted average is not enough. Arbitration must compare every candidate with a baseline, separate evidence from assumptions and show what would change the decision.
What you receive
A reproducible and contestable arbitration
The committee receives the recommendation and everything required to reproduce, challenge or revise it when the context changes.
- Versioned frame
- Decision, baseline, axes, anchors, weights and decision gates
- Evidence register
- Source, date, scope, reliability and author for every judgement
- Robustness analysis
- Intervals, sensitivities, switching values and ranking conflicts
- Decision log
- Decision, rationale, conditions, authority, due date and review trigger
The protocol
From question to ranking, in the right order
The score is never the starting point. The question, the option without a new AI system and the constraints are fixed before trade-offs are examined.
Examine the method in detail
The domainsFive control lenses, criteria built for the decision
The five domains prevent blind spots; they prescribe neither five scores nor five fixed weights. Each engagement derives criteria that distinguish the options without counting the same effect twice.
| Domain | Question examined | Double count to remove |
|---|---|---|
| Useful effects | Which attributable change is expected against the baseline? | Do not count the same gain under value, adoption and feasibility |
| Data | Can the required data be used for the target population? | Separate measured quality, rights and the cost of remediation |
| Feasibility | Does the chain hold under target and degraded conditions? | Do not turn an uncontrolled risk into mere technical effort |
| Operating capability | Are the process, oversight and skills practicable? | Distinguish willingness to adopt from demonstrated capability |
| Risks and impacts | Which negative effects remain and who can stop them? | An unacceptable limit becomes a gate, not a low score |
Useful effects
- Question examined
- Which attributable change is expected against the baseline?
- Double count to remove
- Do not count the same gain under value, adoption and feasibility
Data
- Question examined
- Can the required data be used for the target population?
- Double count to remove
- Separate measured quality, rights and the cost of remediation
Feasibility
- Question examined
- Does the chain hold under target and degraded conditions?
- Double count to remove
- Do not turn an uncontrolled risk into mere technical effort
Operating capability
- Question examined
- Are the process, oversight and skills practicable?
- Double count to remove
- Distinguish willingness to adopt from demonstrated capability
Risks and impacts
- Question examined
- Which negative effects remain and who can stop them?
- Double count to remove
- An unacceptable limit becomes a gate, not a low score
Decision gatesWhat is unacceptable cannot be compensated
The authority sets its risk tolerance and non-compensatory conditions before seeing the scores. These examples are adapted to the portfolio’s context, sector and legal regime.
| Non-compensatory condition | Effect | Evidence required to reopen |
|---|---|---|
| Prohibited use or residual impact outside tolerance | Stop | Revised qualification and risk brought below the approved threshold |
| Data without sufficient rights, coverage or quality | Defer | Rights, population, quality and lineage established |
| No owner, human oversight or fallback mode | Defer | Mandated owner and tested operating controls |
| Performance not demonstrated under target conditions | Reframe | Test protocol, baseline and reproducible results |
Prohibited use or residual impact outside tolerance
- Effect
- Stop
- Evidence required to reopen
- Revised qualification and risk brought below the approved threshold
Data without sufficient rights, coverage or quality
- Effect
- Defer
- Evidence required to reopen
- Rights, population, quality and lineage established
No owner, human oversight or fallback mode
- Effect
- Defer
- Evidence required to reopen
- Mandated owner and tested operating controls
Performance not demonstrated under target conditions
- Effect
- Reframe
- Evidence required to reopen
- Test protocol, baseline and reproducible results
RobustnessA recommendation must survive plausible assumptions
The pivot scenario is recalculated across documented bounds. A recommendation that changes after the first reasonable movement in a score or weight is not presented as a priority.
- Score uncertainty
- Calculate lower and upper bounds from the available evidence
- Weight sensitivity
- Vary each weight within the range accepted by the authority
- Switching value
- Identify the exact assumption that reverses the ranking or decision
- Stability
- Classify the result as robust, conditional or indeterminate
Three examined casesA relative ranking, never a universal merit score
These values are entirely synthetic. They use an illustrative profile to verify the calculation; a live decision adds value functions, gates, sources, intervals and weight scenarios.
| Use case | Impact | Data | Feasibility | Adoption | Controls | Index | Reading |
|---|---|---|---|---|---|---|---|
| A — Sales-call transcription and CRM update | 70 | 85 | 80 | 70 | 65 | 74.25 | Rank 1, still conditional |
| B — Legal contract review assistant | 90 | 35 | 70 | 55 | 30 | 60.75 | Rank 2, evidence to strengthen |
| C — Generative design for a production line | 55 | 40 | 35 | 50 | 45 | 45.75 | Rank 3, dominated in this scenario |
A — Sales-call transcription and CRM update
- Impact
- 70
- Data
- 85
- Feasibility
- 80
- Adoption
- 70
- Controls
- 65
- Index
- 74.25
- Reading
- Rank 1, still conditional
B — Legal contract review assistant
- Impact
- 90
- Data
- 35
- Feasibility
- 70
- Adoption
- 55
- Controls
- 30
- Index
- 60.75
- Reading
- Rank 2, evidence to strengthen
C — Generative design for a production line
- Impact
- 55
- Data
- 40
- Feasibility
- 35
- Adoption
- 50
- Controls
- 45
- Index
- 45.75
- Reading
- Rank 3, dominated in this scenario
Sources and references
- UK Government — practical multi-criteria analysis guidance (new tab)
Value functions, swing weighting, decision conferencing, sensitivity and ranking robustness.
- UK Government — Green Book 2026 (new tab)
Swing weighting, sensitivity analysis, switching values and the warning against simple weighted scoring.
- NIST AI RMF 1.0 (new tab)
Govern, map, measure and manage the risks of an AI system.
- ISO/IEC 42005:2025 (new tab)
Assessing the impacts of an AI system on individuals, groups and society.
Frequently asked questions
How can use cases with different effects be compared?
The mandate fixes one decision question, one baseline and the resources in contention. Raw performances are then translated through explicit value functions before weighting.
Why not use a weighted average directly?
Because a total can conceal an unacceptable condition, double-counting or a trade-off the authority does not accept. Gates, scales and model admissibility are checked before calculation.
Do the five domains have fixed weights?
No. They are a coverage control. Criteria and ranges are constructed for the decision; weights then compare the value of changes of state through swing weighting.
Does first rank automatically trigger funding?
No. Rank is relative to admissible options and the versioned frame. The authority receives possible reversals, uncertainty and limitations, then makes and signs the decision.
First conversation
Which decision must your portfolio make possible?
State the candidates, the authority deciding and the resources in contention. We will specify the protocol, evidence and robustness tests required.