1. Frame the board decision
Define the capability, business outcome, decision horizon, risk appetite and accountable owner before comparing routes.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a board decision charter.
The principal failure occurs when an AI initiative starts with a preferred vendor or acquisition target. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for frame the board decision should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
2. Define the agentic use case
Map goals, tools, data, permissions, memory, hand-offs, human oversight and failure consequences.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an agent operating model.
The principal failure occurs when agentic AI is treated as a generic productivity layer. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for define the agentic use case should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
3. Set the strategic clock
Estimate the latest useful deployment date and the cost of delay by customer, process and competitive response.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a time-to-capability model.
The principal failure occurs when speed has value without an expiry curve. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for set the strategic clock should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
4. Map capability layers
Separate models, orchestration, data, workflow, integration, evaluation, security and change management.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a capability-stack map.
The principal failure occurs when one provider supplies a complete defensible capability. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for map capability layers should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
5. Identify the control perimeter
Specify decisions, data, intellectual property, infrastructure and customer relationships that require ownership.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a control-rights register.
The principal failure occurs when control is equated with owning every component. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for identify the control perimeter should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
6. Measure capability maturity
Test prototype evidence, production reliability, domain performance, adoption and operating economics.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a maturity evidence scorecard.
The principal failure occurs when a demonstration establishes production readiness. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for measure capability maturity should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
7. Test internal build capacity
Assess product leadership, engineering, domain experts, data readiness, safety, procurement and operations.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a build-readiness assessment.
The principal failure occurs when headcount availability equals build capability. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test internal build capacity should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
8. Estimate build economics
Model talent, compute, data preparation, integration, evaluation, assurance, maintenance and opportunity cost.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a build total-cost model.
The principal failure occurs when prototype cost represents lifecycle cost. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for estimate build economics should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
Table 1. Route economics
| Route | Primary cost | Hidden cost |
|---|---|---|
| build | talent and compute | delay and maintenance |
| buy | price and integration | retention and dependencies |
| partner | usage and implementation | lock-in and egress |
Illustrative analytical design; company-specific evidence and professional advice govern.
Values are illustrative evidence indices and require company-specific support.
9. Assess build defensibility
Test proprietary data, workflow integration, feedback loops, distribution and organisational learning.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a defensibility map.
The principal failure occurs when custom code automatically creates advantage. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for assess build defensibility should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
10. Assess build execution risk
Quantify hiring delay, architecture churn, model obsolescence, security defects and adoption shortfall.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a build risk register.
The principal failure occurs when internal delivery removes supplier risk. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for assess build execution risk should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
11. Define the acquisition thesis
Specify scarce capability, people, data, customers, licences and time advantage sought.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an acquisition thesis.
The principal failure occurs when an AI label constitutes strategic fit. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for define the acquisition thesis should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
12. Map target dependency
Identify foundation-model, cloud, chip, open-source, data and key-person dependencies.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a dependency diligence map.
The principal failure occurs when the target owns the whole technology stack. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for map target dependency should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
13. Test technical quality
Review architecture, evaluations, reliability, latency, observability, security and technical debt.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a technical diligence protocol.
The principal failure occurs when benchmark claims transfer directly to the buyer environment. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test technical quality should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
14. Test data rights
Trace provenance, consent, licences, retention, localisation, derived data and training rights.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a data-rights chain.
The principal failure occurs when database possession proves lawful reusable rights. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test data rights should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
15. Test model and IP rights
Review weights, code, prompts, fine-tuning assets, open-source obligations, patents and contractor assignments.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an AI IP schedule.
The principal failure occurs when registered corporate ownership captures all AI intellectual property. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test model and ip rights should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
16. Test safety and assurance
Evaluate harmful outputs, tool misuse, prompt injection, leakage, autonomy, monitoring and incident response.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an agentic safety case.
The principal failure occurs when human review is an adequate universal control. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test safety and assurance should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
Table 2. AI diligence layers
| Layer | Evidence | Decision |
|---|---|---|
| technology | evaluations and architecture | fitness |
| data | provenance and rights | lawful use |
| safety | tests and controls | residual risk |
| regulation | role and obligations | permission |
Illustrative analytical design; company-specific evidence and professional advice govern.
Values are illustrative evidence indices and require company-specific support.
17. Test regulatory exposure
Map provider, deployer and importer roles, risk categories, transparency duties and sector rules.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a regulatory obligations map.
The principal failure occurs when compliance follows the seller after completion. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test regulatory exposure should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
18. Test competition exposure
Assess compute access, cloud commitments, exclusivity, information rights, talent and ecosystem foreclosure.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a competition-risk memorandum.
The principal failure occurs when a minority investment or partnership avoids merger scrutiny. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test competition exposure should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
19. Value the acquisition
Separate proven revenue, capability option value, time saved, replacement cost and defensible synergies.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a capability-adjusted valuation.
The principal failure occurs when strategic urgency justifies an unbounded premium. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for value the acquisition should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
20. Model integration reality
Price platform migration, identity, data, evaluation, security, product integration and customer continuity.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an integration cost bridge.
The principal failure occurs when AI assets integrate like conventional software. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for model integration reality should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
21. Protect critical talent
Identify capability holders, incentives, research autonomy, culture and retention periods.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a talent-continuity plan.
The principal failure occurs when standard retention awards preserve tacit knowledge. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for protect critical talent should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
22. Design acquisition terms
Align earn-outs, holdbacks, IP warranties, data indemnities, regulatory conditions and founder commitments.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an AI transaction term sheet.
The principal failure occurs when conventional technology warranties cover agentic risk. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for design acquisition terms should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
23. Define the partnership thesis
Specify the capability gap, partner contribution, duration, reversibility and learning objective.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a partnership charter.
The principal failure occurs when partnership is a low-commitment default. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for define the partnership thesis should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
24. Map partnership economics
Model usage, minimum spend, revenue share, implementation, support, egress and renegotiation.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a partnership economics model.
The principal failure occurs when headline API price captures economic dependence. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for map partnership economics should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
Table 3. Partnership control
| Right | Risk | Protection |
|---|---|---|
| data use | unintended reuse | purpose limits |
| model change | performance drift | version control |
| termination | operational failure | transition support |
| IP | value leakage | foreground allocation |
Illustrative analytical design; company-specific evidence and professional advice govern.
Values are illustrative evidence indices and require company-specific support.
25. Map lock-in
Assess proprietary interfaces, fine-tuning, data gravity, workflow embedding, cloud commitments and switching time.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a lock-in stress test.
The principal failure occurs when multi-vendor language assures practical portability. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for map lock-in should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
26. Allocate rights and control
Define foreground IP, improvements, data use, audit, subcontracting, model change and termination assistance.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a rights-and-control matrix.
The principal failure occurs when a standard licence allocates strategic rights. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for allocate rights and control should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
27. Test partner resilience
Assess funding, capacity, roadmap, service continuity, security incidents and ownership change.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a partner resilience review.
The principal failure occurs when market prominence guarantees continuity. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for test partner resilience should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
28. Design exit portability
Require data export, prompt and evaluation assets, replacement interfaces, documentation and transition capacity.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a portability plan.
The principal failure occurs when contract termination creates operational exit. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for design exit portability should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
29. Compare routes consistently
Score build, buy and partner against speed, control, economics, risk, learning, flexibility and value.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a route-comparison scorecard.
The principal failure occurs when each route is assessed using a different business case. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for compare routes consistently should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
30. Use real options
Value staged pilots, minority investment, call rights, licensing, acquihire and later acquisition.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an option-architecture map.
The principal failure occurs when the board must choose one permanent route immediately. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for use real options should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
31. Design hybrid routes
Combine owned workflow, partnered models, acquired talent and replaceable infrastructure.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a modular sourcing architecture.
The principal failure occurs when build, buy and partner are mutually exclusive. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for design hybrid routes should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
32. Run scenario tests
Test model commoditisation, regulation, compute scarcity, vendor failure, cyber incident and rapid demand growth.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a route scenario model.
The principal failure occurs when the base case represents the decision distribution. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for run scenario tests should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
Table 4. Scenario comparison
| Scenario | Route pressure | Board response |
|---|---|---|
| commoditisation | buy premium erodes | stage commitment |
| vendor failure | partner continuity falls | portability |
| regulation | assurance cost rises | control evidence |
| demand surge | capacity tightens | reserved access |
Illustrative analytical design; company-specific evidence and professional advice govern.
Values are illustrative evidence indices and require company-specific support.
33. Set decision thresholds
Define minimum performance, maximum lock-in, payback, control and safety requirements.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a non-negotiable gate register.
The principal failure occurs when a weighted score can compensate for a failed critical control. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for set decision thresholds should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
34. Sequence evidence
Use discovery, sandbox, controlled pilot, diligence and commitment gates.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a staged evidence plan.
The principal failure occurs when the organisation commits capital before learning cheaply. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for sequence evidence should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
35. Govern model changes
Require version testing, approval, rollback, documentation and customer communication.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a model-change control.
The principal failure occurs when supplier model updates are routine software patches. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for govern model changes should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
36. Govern agent permissions
Apply least privilege, tool allowlists, transaction limits, human escalation and kill controls.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an agent authority matrix.
The principal failure occurs when an agent should receive broad access to maximise utility. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for govern agent permissions should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
37. Measure realised value
Track adoption, task completion, error, cycle time, revenue, cost, risk and learning.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a benefits-realisation dashboard.
The principal failure occurs when usage volume establishes strategic return. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for measure realised value should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
38. Create board reporting
Report capability progress, route economics, incidents, dependencies, regulation and option value.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is an AI board pack.
The principal failure occurs when technical milestones alone support board oversight. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for create board reporting should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
39. Define operating accountability
Assign product, technology, data, security, legal, risk, finance and business ownership.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a cross-functional accountability map.
The principal failure occurs when accountability sits solely with the AI team. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for define operating accountability should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
40. Close through evidence gates
Approve the route only when capability, economics, rights, safety, portability, integration and ownership are evidenced.
The controlled record should identify the relevant customer, revenue stream, contract, cohort, relationship owner, commercial dependency, evidence source, baseline, trend, scenario, control, exception, accountable executive, deadline and approval. The immediate deliverable is a board route certificate.
The principal failure occurs when strategic excitement overrides unresolved dependencies. Reviewers should connect the evidence to revenue durability, gross margin, cash conversion, renewal probability, switching behaviour, counterparty credit and transaction value; distinguish contractual protection from observed customer behaviour; and test whether the conclusion survives a downside scenario.
The decision pack for close through evidence gates should state the commercial question, measurement perimeter, historical evidence, customer-specific facts, forecast logic, sensitivity, management action, buyer implication, residual uncertainty and next gate. Material exceptions should flow into the valuation model, quality-of-earnings work, diligence room, sale-process narrative, transaction protections and board reporting.
Table 5. Board gates
| Gate | Minimum evidence | Owner |
|---|---|---|
| capability | production evaluation | product |
| economics | lifecycle model | finance |
| rights | controlled schedule | legal |
| safety | approved assurance case | risk |
Illustrative analytical design; company-specific evidence and professional advice govern.
Values are illustrative evidence indices and require company-specific support.
References
- National Institute of Standards and Technology, AI Risk Management Framework, https://www.nist.gov/itl/ai-risk-management-framework
- National Institute of Standards and Technology, Generative AI Profile NIST AI 600-1, https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.600-1.pdf
- NIST AI Resource Center, AI RMF resources, https://airc.nist.gov/
- European Commission, AI Act regulatory framework, https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
- European Commission, General-Purpose AI model obligations, https://digital-strategy.ec.europa.eu/en/faqs/guidelines-obligations-general-purpose-ai-providers
- European Commission, General-Purpose AI Models Q&A, https://digital-strategy.ec.europa.eu/en/faqs/general-purpose-ai-models-ai-act-questions-answers
- UK Competition and Markets Authority, AI Foundation Models update paper, https://www.gov.uk/government/publications/ai-foundation-models-update-paper
- UK Competition and Markets Authority, AI Foundation Models technical update report, https://assets.publishing.service.gov.uk/media/661e5a4c7469198185bd3d62/AI_Foundation_Models_technical_update_report.pdf
- UK Competition and Markets Authority, AI strategic update, https://www.gov.uk/government/publications/cma-ai-strategic-update/cma-ai-strategic-update
- UK Competition and Markets Authority, AI partnerships and other arrangements, https://www.gov.uk/government/news/cma-seeks-views-on-ai-partnerships-and-other-arrangements
- US Federal Trade Commission, AI Partnerships and Investments 6(b) Study, https://www.ftc.gov/reports/ftc-staff-report-ai-partnerships-investments-6b-study
- US Federal Trade Commission, Partnerships Between Cloud Service Providers and AI Developers, https://www.ftc.gov/system/files/ftc_gov/pdf/p246201_aipartnerships6breport_redacted.pdf
- US Department of Justice and Federal Trade Commission, 2023 Merger Guidelines, https://www.justice.gov/atr/2023-merger-guidelines
- UK Competition and Markets Authority, Merger Assessment Guidelines, https://www.gov.uk/government/publications/merger-assessment-guidelines
- Organisation for Economic Co-operation and Development, OECD AI Principles, https://oecd.ai/en/ai-principles
- Organisation for Economic Co-operation and Development, Framework for the Classification of AI Systems, https://www.oecd.org/en/publications/oecd-framework-for-the-classification-of-ai-systems_cb6d9eca-en.html
- International Organization for Standardization, ISO IEC 42001 AI management systems, https://www.iso.org/standard/81230.html
- International Organization for Standardization, ISO IEC 23894 AI risk management, https://www.iso.org/standard/77304.html
- UK National Cyber Security Centre, Guidelines for Secure AI System Development, https://www.ncsc.gov.uk/collection/guidelines-secure-ai-system-development
- Cybersecurity and Infrastructure Security Agency, Guidelines for Secure AI System Development, https://www.cisa.gov/resources-tools/resources/guidelines-secure-ai-system-development
- UK Government, AI regulation white paper, https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach
- US Office of Management and Budget, M-24-10 Advancing Governance Innovation and Risk Management for Agency Use of AI, https://www.whitehouse.gov/wp-content/uploads/2024/03/M-24-10-Advancing-Governance-Innovation-and-Risk-Management-for-Agency-Use-of-Artificial-Intelligence.pdf
- International Valuation Standards Council, International Valuation Standards, https://ivsc.org/standards/
- IFRS Foundation, IFRS 3 Business Combinations, https://www.ifrs.org/issued-standards/list-of-standards/ifrs-3-business-combinations/
- IFRS Foundation, IAS 38 Intangible Assets, https://www.ifrs.org/issued-standards/list-of-standards/ias-38-intangible-assets/
- IFRS Foundation, IFRS 13 Fair Value Measurement, https://www.ifrs.org/issued-standards/list-of-standards/ifrs-13-fair-value-measurement/
