Digital Strategy Playbook

Nine tenets, one map and five choices carry the whole method.

The tenets, the three-phase method, the five choices, who decides, and what the worked example showed. Read this if you read nothing else: three phases, five strategic choices, twelve steps, four decision gates and twenty-five working tools.

This is a practical method for business-led and digital-led choices on digital, data and AI, and for governing their delivery. It supports two directions: a business strategy that needs digital capabilities, and digital possibilities that reshape business strategy. In both, the aim is a set of choices the business can defend and fund.

Nine tenets govern every route through the method.

Each tenet names a decision made explicitly. Later steps, gates and tools cite them by number.

Tenets 1 to 3 · What strategy must answer

Tenets 4 to 6 · How choices are made

Tenets 7 to 9 · What must actually change

Tenet 1 admits digitally originated opportunities as well as business questions. Tenet 5 requires explicit decision rights configured to the enterprise, rather than one universal structure.

One map: three phases, twelve steps, four gates, five choices.

Vision → Strategy → Blueprint organizes the steps; later evidence can reopen earlier decisions. The map is the navigation model for the whole playbook: the three phases, the twelve numbered steps, the four executive gates, the diagnosis synthesis checkpoint, the five strategic choices and the feedback routes.

Gates and the checkpoint

• Mandate gate, after Step 2.
• Synthesis checkpoint, after Step 5. It is a working review, not a fifth gate.
• Direction gate, after Step 7.
• Coherence gate, after Step 9.
• Commitment gate, after Step 10.

Where the five strategic choices are developed and confirmed

The choices run through the steps; they are not extra steps. Developed means the choice is worked or held provisionally; confirmed means it is decided or operated.

• 1. Winning aspiration: developed in Steps 2 to 6, confirmed at Step 7.
• 2. Where to play: developed in Steps 4 to 6, confirmed at Step 7.
• 3. How to win: developed in Step 6, confirmed in Steps 7 and 8.
• 4. Required capabilities: developed in Steps 5 to 8, confirmed in Steps 9 and 10.
• 5. Management systems: developed in Step 9, confirmed and operated in Steps 10 to 12.

Phases organize strategic thinking, not a waterfall. A common misreading treats the three phases as a replacement for the twelve steps; they group the steps, they do not replace them. Choices adapted from A.G. Lafley and R. Martin, Playing to Win (2013); the mapping to the twelve steps is this playbook’s adaptation.

Five choices give leaders one language across the twelve steps.

Each choice is developed in some steps and confirmed at a gate; none is an extra step.

For mission-led organizations, read winning as superior mission outcomes and public value relative to alternatives. A useful test for an executive team: can each member state the answer to all five questions in one sentence each, and do the answers agree?

Source: the five questions are from A.G. Lafley and R.L. Martin, Playing to Win: How Strategy Really Works (Harvard Business Review Press, 2013). The step, tool and gate mapping, and the noncommercial reading, are adaptations developed for this playbook.

Every route ends in the same eight strategy deliverables.

This minimum viable strategy package is what the sponsor receives. Tools are means to these outputs; use a tool only where it supplies decision evidence the package needs. The route changes evidence depth, not the list, and deliverables mature from provisional to approved across the gates.

Each gate releases only what its evidence supports.

Approval is proportionate to exposure. A conditional direction does not authorize full implementation. Commitment grows in stages: discovery, then capped validation, then tranches, each tied to evidence.

Decision rights are explicit, whatever the organizational design.

This is a reference pattern for a centralized functional design. Other enterprises use federated, product-based or hybrid patterns. Every configuration needs one outcome owner, named approvers, and blocking rights that cite a policy and carry an escalation route.

Premium uptime won; the platform failed its payback test.

Meridian Industrial Services is a fictional 20-site business; all figures are illustrative. The full case, with every record, is in Part four. The example shows the method changing its own first hypothesis.

USD millions, 20 sites, three-year horizon, 10 percent discount rate. Figures exclude taxes, financing, inflation, terminal value and the $0.05m sunk discovery spend.

Five decisions this playbook asks of leaders.

The method only works if leaders make these choices explicitly, before work and money move. It is worth asking who in the organization currently makes each one, and whether it is written down.

How to use this playbook

The playbook is a decision sequence with the tools attached, not reading material. Begin with the business value question or a credible digital opportunity, never with a predetermined technology solution, then pick the route that matches the consequence and reversibility of the decision.

What is in this playbook

An executive summary, five parts and the annexures. Leaders usually need the executive summary and Part four; practitioners will live in Parts three and five.

• Executive summary: the tenets, the three-phase method, the five choices, who decides, and what the worked example showed.
• Part one, Strategy in context: what strategy and digital strategy are, Digital’s purpose, and the economics that set digital and AI apart.
• Part two, The integrated method: twelve steps, four decision gates, and the digital lens for each step.
• Part three, The twelve steps: method, questions, traps and completion criteria for every step, with decision rights at Step 9.
• Part four, Worked examples: Meridian (business-led) completes every step; Halden (digital-led) and Ardent (technology renewal) follow.
• Part five, Practical tools: twenty-five tools linked by one identifier chain; thirteen templates in this playbook, all 25 in the toolkit workbook.
• Annexures: methods mapped to the steps, glossary, references, the Strategic Alignment Model, and from choice to execution.

Choose the depth the decision deserves.

Begin with the decision to be made. Select a route, reuse trustworthy evidence and agree who can authorize discovery, validation and commitment.

Route choice depends on consequence, reversibility, risk, complexity, available evidence and uncertainty. The indicative periods are elapsed time to a decision, and they exclude longer tests. Every route still covers all four decision purposes.

Nine factors select the route, and the strictest one can override.

Score each factor before Step 1 and record the route in Tool 1. Use the route most factors indicate, unless reversibility or regulatory exposure calls for a stricter one. A high rating on irreversibility, regulatory exposure or investment exposure moves the work to at least the standard route. Record the reasoning whenever the chosen route differs from what the ratings suggest.

Each route sets minimum depth, challenge and authority.

Routes differ in evidence and governance, not only in duration. An initial decision is not validation, and neither is funding or benefit. Four clocks run separately: time to an initial decision, to validation evidence, to authorized investment, and to benefits. A rapid route can reach a conditional direction quickly; it cannot validate customer behavior or benefits in that time.

Indicative times. Field tests, implementation and benefit realization run on their own clocks and are planned in Steps 7, 10 and 12, not compressed to fit the workshop sequence.

Each route may compress some work, never the controls.

The minimum package states who takes part, what may be shortened, what cannot be omitted and what the route can authorize. Read each column top to bottom: when the route fits, who must take part, which forums decide, what may be shortened, what can never be omitted, when to escalate, and what the route can and cannot authorize.

Workshop duration is not elapsed time: evidence gathering and validation run between sessions, as the rapid route section sets out. A route can be escalated at any point; it is never downgraded after a gate without recording why.

Rapid route: an authorized decision in ten days.

For a bounded, reversible choice on reliable evidence. Ten days reaches an initial authorized decision, not completed testing. Every decision purpose still runs; meetings are combined only under explicit authority.

The rapid route still needs at least two materially different options plus the continuation baseline, and any exception is recorded. It does not reach completed testing or benefits in ten days.

Workshop time is not elapsed time.

Ten working days on the rapid route reaches an initial authorized decision on existing evidence, not completed tests, delivery or benefits. Scale depth to consequence, reversibility, risk exposure, complexity, evidence availability and uncertainty.

The step cards show facilitated workshop time only. Preparation, evidence collection and field validation run between workshops and set the elapsed time, so plan the calendar around evidence, not meetings. The standard route is shown; durations are indicative.

Strategy is a diagnosis and a set of linked choices.

Strategy is a coherent set of consequential choices that addresses a diagnosed challenge and directs scarce resources. Analysis informs those choices, plans translate them, and operations deliver them.

What strategy is, and is not.

The pairs separate real strategy from the artifacts most often mistaken for it. Use them to challenge documents that call themselves strategies: the test is whether there is a diagnosis, linked choices, explicit trade-offs and a theory of value.

Analysis informs choices, plans translate them, operations deliver them. None of the three is a substitute for a choice.

Choices connect to sustained delivery.

Choices create value only when the business model, operating model and resources carry them through. Evidence can reopen any of the three.

• Strategic choices: where and how to win.
• Business model: create and capture value.
• Operating model: capabilities and work.
• Portfolio: resources and delivery.
• Operations: adoption and sustainment.
• Realized value: evidence and adaptation.

The chain shows how a choice becomes delivery: strategic choices set where and how to win, the business model turns that into value created and captured, the operating model builds the capabilities to deliver it, and the portfolio funds the work. Operations sustain adoption, and the value actually realized becomes the evidence that can reopen any step before it. That is the purpose of the return path in the method.

Three layers, designed together but owned separately.

Each answers a different question. Collapsing them hides decisions, assumptions and ownership. Keep the artifacts separate but design them together.

Tactical work turns strategic choices into operations.

Tactical management is the bridge. It converts strategic choices into portfolios, roadmaps, funding, capacity, sequencing and accountabilities that operations can absorb. Strategies fail here more often than in the choice itself.

Means and ends link purpose to every KPI.

Every means serves a named end at the same level. Purpose and aspiration stand in for the model’s mission and vision; “value driver” keeps its economic meaning. Goals set at Step 2 are provisional: they are confirmed or revised once the diagnosis is complete.

Source: Business Motivation Model, Object Management Group, v1.3 (2015), adapted: purpose and aspiration replace mission and vision; influencers are the model’s drivers.

Every level of strategy runs the same twelve steps.

The same twelve steps work at corporate, business-unit, functional and transformation level. What changes by level is the evidence, the economics and the decision forum, not the method.

Twelve predictable failure modes, each with a corrective.

Each failure mode has a corrective a decision forum can apply. Six are failures of strategy in general; six are specific to digital, and digital failures grow as AI moves from pilots into everyday work. A useful exercise is to name the two failure modes seen most often in your own organization.

Digital strategy serves business strategy, and can reshape it.

Business strategy decides where to play and how to win. Digital strategy decides how digital, data, AI and technology enable those choices, and where they open choices the business had not considered.

Either way, accountable business leadership validates and authorizes the enterprise value proposition.

A digital strategy that cannot name the business value it serves or creates is a technology plan, not a strategy.

Source: Strategy Methods toolkit (2026).

Three strategy products share one twelve-step method.

Name the product at Step 1. They are perspectives on one strategy, not three separate processes; most work has a primary product and touches the others. They differ in the question answered, the outputs and who decides.

Source: developed for this playbook. The four types of digital strategy, set out later in this part, show which domain drives each change.

Digital creates value only when the business does.

Digital’s aspiration names Digital as its subject but points at a business outcome. Trust is part of it, because growth that stakeholders cannot rely on does not last. The statements are an example to adapt.

The 4Es

Nine reference objectives for a functional Digital/IT strategy.

Objectives 1 to 3 describe what Digital delivers. Objectives 4 to 9 describe how Digital must operate to deliver them.

Seven principles test the objectives and their measures.

Use the principles to design and challenge the measure set. Coverage is tested across the strategy and portfolio, not forced onto every individual objective.

Which domain leads defines four types of digital strategy.

Each alignment perspective of the Strategic Alignment Model runs influence in a different direction: one domain drives the change, a second carries it, and a third has to adapt. Strategy execution and technology transformation start from business strategy; competitive potential and service level start from digital strategy. Name the type at Step 1, tag each initiative in Step 11, and expect most strategies to mix types across the portfolio.

Source: Henderson, J. C. and Venkatraman, N. (1993), Strategic alignment, IBM Systems Journal, 32(1). See the Strategic Alignment Model method section in the annexures.

Five economic features change what each step must test.

None holds universally. Each is a condition to test when digital, data or AI sits on the crux; the technology models that follow compare how the economics differ.

Marginal cost, for example, is low for many digital services, yet cloud consumption and AI inference can make it material. Treat each feature as a condition to test, not an assumption.

Economics differ by technology model, so test the one in play.

Use this at Steps 6 to 8 to set the cost lines and the decisive economic test for each option. Hybrid options combine rows. Agentic AI, for example, is judged on fully loaded cost per completed business task, not on license price.

Data creates advantage through a combination of factors.

Assess eight factors in Step 5 with Tool 5. No single factor is required in every case: widely available data can create advantage when combined with superior capabilities, integration and closed-loop learning. Advantage does not come from uniqueness alone.

Each level of AI autonomy needs its own decision rights.

AI agents move work from people to systems. Step 9 sets the autonomy level for each use case, and Tool 23 records it with its controls.

AI is judged on strategy and economics, then governed.

Assess every material AI option from four perspectives, alongside its non-AI alternative. Tool 23 records the result.

AI is governed across its whole life, in proportion to risk.

One Tool 23 card follows each predictive, generative or agentic use case from opportunity to retirement. More autonomy is not better by default. Model approval and deployment approval are separate decisions, any change of model or provider triggers reassessment, and cost is measured per successfully completed business task.

Autonomy levels as defined in the autonomy table at the start of this section. Terminology is informed by ISO/IEC 42001 and the NIST AI Risk Management Framework; using this playbook does not establish conformity with either.

A successful demonstration is not evidence for scale.

AI is evaluated in three settings, and each setting can authorize only the next step. Laboratory evaluation tests capability on representative held-out cases; pilot performance tests the real workflow with real users and support; routine operational performance tests whether results hold at volume, without exceptional support, over time. A laboratory or demonstration result never justifies enterprise deployment on its own.

Every material AI proposal also names its value driver, non-AI alternative, data rights, permitted actions, accountable person, provider exit and conditions for raising scope or autonomy (Tool 23). Informed by the NIST AI RMF and ISO/IEC 42001; no conformity is implied.

The method moves from mandate to learning.

The integrated method is twelve steps, four gates and a digital lens for each step. It alternates disciplined divergence with explicit convergence, and four decision gates prevent teams from advancing when authority, direction, coherence or commitment remain unresolved.

The twelve steps apply unchanged to a digital strategy. What changes is the evidence each step needs and the methods that produce it. Teams explore widely, then decide explicitly.

Steps 1 to 6 anchor digital in business value.

For the first six steps, the emphasis is on anchoring digital in business value and on testing capability on three separate scales: maturity, importance and advantage.

Source: Strategy Methods toolkit (2026).

Steps 7 to 12 fund digital in waves, on evidence.

Later steps turn digital choices into a governed portfolio. Minimum viable foundations come first, and each wave is funded on the evidence of the one before.

Source: Strategy Methods toolkit (2026).

Steps 1 to 6: owner, decision and exit test.

Use these practitioner step cards to brief step owners. Each row gives the decision the step supports, the accountable owner and the exit test. Inputs, participants, questions and traps sit in each step’s own section in Part three. A step is complete only when its exit test is met, not when its workshop has run.

D1 to D8 are the eight minimum strategy deliverables every route ends in. On the rapid route the exit tests are the same; the evidence depth is lighter.

Steps 7 to 12: owner, decision and exit test.

From choice to realized value. Each later step tests the earlier choice, and a failed or inconclusive exit test returns work to the step that made the assumption.

Step 12 reviews are not gates. They use the five realization decisions (continue, scale, adapt, pause, exit) and can reopen Steps 6 to 9.

Evidence carries a decision only if it survives six tests.

Six evidence tests set the bar for every material claim, four gates each make a different leadership decision, and two vocabularies keep test results apart from forum decisions. A tenet map shows where each principle is applied and recorded.

Six tests every material claim must pass.

This is tenet 6 in practice. Apply the tests before a claim enters a gate paper, and apply them proportionately: the more consequential and irreversible the decision, the higher the bar.

Four gates, four different decisions.

Each gate asks a different question and leaves a different record. Before any gate meeting, make sure the forum knows which decision it is being asked to make.

Five gate decisions and four test results, each with one meaning.

Forums decide; tests inform. Keep the two vocabularies apart: analysts and testers produce test results, and forums make decisions. A Pass result never authorizes anything by itself, Inconclusive is not Fail, and every gate records exactly one of the five statuses in Tool 13, with owners and expiry dates for any conditions.

Step 12 reviews are not gates: they use the five realization decisions (continue, scale, adapt, pause, exit).

Each tenet has a home in the steps, gates and tools.

Use this map to find where each of the nine tenets is applied and which tool records it. It doubles as an audit checklist: if a tenet has no evidence in its tool, the gate should not pass.

Match the working posture to the environment.

Predictability sets the working posture, not the method. In uncertain environments, use competing hypotheses and small reversible commitments; in predictable ones, commit within a tested envelope.

Step 1 sets the scope of the digital strategy, the business decision it serves and who may authorize it.

Without that anchor, digital work drifts into a technology plan. Part three sets out the twelve steps: each states its purpose, decision owner, inputs and outputs, the method, the questions that matter, the traps that recur, and the criteria that say it is complete.

Step 1 · Define the strategic mandate & scope

Turn a broad request for a strategy into a bounded decision with authority, evidence standards and a realistic process. Record the strategy product, whether the work is business-led or digital-led, and the alignment perspective. A digital-led opportunity can enter here without an artificial existing business mandate, but a business owner must own the value question.

Workshop time: 90 minutes.

Completion checklist

• 1. The unit of strategy is explicit
• 2. The decision question is written
• 3. The horizon and boundaries are clear
• 4. Decision rights are named
• 5. Evidence standards are agreed
• 6. Stakeholders and challengers are included
• 7. The timetable ends in a real decision

Digital method, questions and traps

Use the questions to test whether the request is really a solution looking for a problem. Agree the governance pattern and who can authorize discovery before any analysis starts.

Digital method

• 1. Name the business value question, strategic challenge or digitally enabled opportunity, and its value drivers.
• 2. Set scope: enterprise, unit, product, platform or function, and any IT and OT boundary.
• 3. Write the decision needed now and why digital, data or AI makes it urgent.
• 4. Confirm decision rights for outcomes, rules, delivery and risk, in the governance pattern that fits (see the reference patterns in Step 9).
• 5. Record the risk appetite, AI policy and protected platform allocation as constraints.
• 6. Agree cadence, decision dates and the conditions that would stop or reframe the work.
• 7. Select the strategy product and entry point (the three strategy products in Part one) and the alignment perspective (the four types of digital strategy in Part one).

Step 2 sets enduring purpose, guardrails and a provisional ambition that diagnosis can revise.

Keep purpose and ambition separate from strategy. The ambition stays provisional until the diagnosis is complete.

Step 2 · Establish identity, purpose and ambition

Clarify the enduring context that guides choices, without confusing purpose or ambition with strategy itself. The step sets enduring purpose, values and guardrails, plus a provisional ambition.

Workshop time: half day.

Completion checklist

• 1. Identity is evidenced by behavior
• 2. Purpose names value created
• 3. Ambition is marked provisional until diagnosis
• 4. Guardrails constrain choices
• 5. Ambition is not presented as strategy
• 6. Stakeholders understand the language
• 7. Mandate gate evidence is complete

Digital method, questions and traps

Select only the reference objectives and 4E emphasis that fit this situation. Agree the non-negotiable safety, legal and risk thresholds now, because later scoring must never trade them away.

Digital method

• 1. Restate enduring purpose, values and non-negotiable principles; record current priorities and boundaries.
• 2. Draft a provisional ambition and hypotheses; the checkpoint after Step 5 may revise them.
• 3. For a functional Digital/IT strategy, draft Digital’s purpose (see how digital creates value only when the business does, in Part one) and select relevant reference objectives (the nine reference objectives in Part one).
• 4. Use the 4Es as a lens weighted to the situation; measurable objectives follow in Step 10.
• 5. State what Digital will not do, including the informal veto over demand it gives up.
• 6. Test the draft against the principles and record what remains provisional.

The mandate gate authorizes the process, not a solution.

It releases discovery spend and capacity only. Check that the brief names the strategy product, entry point, outcome owner and evidence standard.

Mandate gate

Authorize the strategy process with a clear problem, scope, intent, authority and evidence standard. The forum is the accountable executive or governing body, and the gate applies tenets 1 and 5.

Gate checklist

• 1. Scope is bounded
• 2. Decision rights are accepted
• 3. Intent guides but does not dictate the answer
• 4. Evidence and challenge standards are credible

Judging evidence and causal explanations

Confidence belongs to a specific claim and decision, not to a whole report. This scale defines what each level requires and how it should change a choice. Low-confidence evidence can justify a test, but never an irreversible commitment.

Step 3 builds the fact base on how processes, platforms and digital services perform today.

Explain the most consequential challenge causally and isolate the crux. Baseline against the 4Es and rate capability maturity separately from strategic importance.

Step 3 · Diagnose performance and the strategic crux

Explain the most consequential challenge or opportunity causally, and isolate the leverage point on which strategy must act. Insist on alternative hypotheses: a diagnosis that has only one explanation has usually not been tested.

Workshop time: half to one day.

Completion checklist

• 1. Baseline measures are reconciled
• 2. Symptoms and causes are separated
• 3. Alternative explanations were considered
• 4. The crux is consequential
• 5. The crux is actionable
• 6. The diagnosis is falsifiable
• 7. Owners accept the facts even if they dispute the response

Digital method, questions and traps

For digital, the fact base covers how processes, platforms and digital services perform today.

Digital method

• 1. Baseline process performance against the 4Es using process mining and operational data.
• 2. Assess platform health: availability, security posture, technical debt and lifecycle risk.
• 3. Split current spend across run, change and innovate, and list benefits already validated.
• 4. Trace symptoms to causes; challenge redesign before automation.
• 5. Rate digital and data capability maturity on the five-level scale, separately from importance.
• 6. Test the four alignment links to find where fit breaks.
• 7. Name the crux and the evidence that would change it.

Step 4 scans the technology, regulatory and competitive forces that change where digital creates value.

Look outside: what customers value, how value and power move, and which external changes alter the problem. Include technology possibilities and value pools, not just competitor activity.

Step 4 · Understand the outside world

Determine what customers value, how value and power move through the market, and which external changes alter the strategic problem.

Workshop time: one to three days plus fieldwork.

Completion checklist

• 1. Customer evidence includes behavior
• 2. Needs and economics are linked
• 3. Industry boundaries are defensible
• 4. Substitutes and complements are included
• 5. Macro drivers have causal implications
• 6. Scenarios expose different decisions
• 7. External assumptions have indicators and owners

Digital method, questions and traps

Look specifically for competitors that compete through data, platforms or AI, and for ecosystem partners who could become rivals.

Digital method

• 1. Scan technology, AI and regulatory trends with PESTEL and S-curves; rank the drivers that matter.
• 2. Test how digital shifts industry structure with Five Forces and platform dynamics.
• 3. Map priority customer journeys and jobs to be done; mark the moments that matter.
• 4. Benchmark rivals’ digital offers, channels and cost to serve.
• 5. Assess suppliers and strategic partners for concentration and exit risk.
• 6. Build two to four scenarios for AI adoption, regulation and cyber threat.

Evidence converges into strategic possibility

External evidence is collected to widen the field of possibility, not to confirm the answer leaders already prefer. This step sits in the outside-in phase of the strategy process, after the ambition has been framed but before any option is chosen. Teams assemble two bodies of external evidence and combine them into a third.

Step 5 tests whether the organization’s capabilities, platforms, data and people can deliver the strategy.

Look inside: resources, capabilities, culture, incentives and constraints. This is where operational maturity, strategic importance and comparative advantage are rated as three separate dimensions.

Step 5 · Understand the inside reality

Identify the resources, capabilities, culture, incentives and constraints that can create advantage or prevent delivery.

Workshop time: one to two days.

Completion checklist

• 1. Capabilities are linked to outcomes
• 2. Claims use comparative evidence
• 3. Distinctive and threshold capabilities are separated
• 4. Culture and incentives are assessed
• 5. Dynamic capability is considered
• 6. Constraints are explicit
• 7. No capability is called distinctive without external evidence

Digital method, questions and traps

Claims of comparative advantage need external evidence such as benchmarks or customer outcomes, not self-assessment.

Digital method

• 1. Build a capability map; rate maturity, strategic importance and comparative advantage separately (see the discipline on rating them separately, after Step 7).
• 2. Map value streams and applications to capabilities to expose duplication and gaps.
• 3. Assess data quality, ownership and access against the Data and AI policy owner’s standards.
• 4. Review architecture, integration and security against the published guardrails.
• 5. Assess skills in Digital and the business, and supplier dependence.
• 6. Claim advantage only on external evidence: cost position, outcomes, benchmarks, scarcity and durability.

Capability evidence and the right to win

A right to win is a comparative claim supported by evidence, not a self-assessment of strength. A capability becomes a defensible claim through a chain of six links. It begins with the resources and routines the organization actually owns, tests whether they combine into something repeatable rather than occasional, and then asks how that performance compares with rivals serving the same customers.

• 1. Resources and routines
• 2. Repeatable capabilities
• 3. Comparative performance
• 4. External benchmarks
• 5. Value and durability
• 6. Right to play or win

Comparison is settled externally, not internally: benchmarks, win rates, cost positions and customer evidence decide whether the capability is genuinely distinctive. Only once value to the customer and durability against imitation are established does the final judgment follow: whether the capability earns a right to play in the market, or a right to win in it.

Diagnosis synthesis checkpoint

After Step 5, confirm, revise, reject or split the crux, and revise the provisional ambition and assumptions where outside-in and inside-out evidence requires. The checkpoint brings the two evidence sets together and can change the provisional ambition. It is a working review, not a fifth approval gate.

Step 6 creates distinct options for how digital, data and AI change the business.

Options should differ in how digital, data and AI change the business, not just in vendor or scope. Standard and enterprise routes need at least three options plus an honest continuation baseline; the rapid route needs at least two. The baseline is a comparator, never one of the options.

Step 6 · Generate strategic options

Create materially different, internally coherent ways to address the crux rather than variations of one preferred answer.

Workshop time: one to two days.

Completion checklist

• 1. Three or more materially different options plus the baseline (two on the rapid route)
• 2. Each option answers where and how to win
• 3. Exclusions are explicit
• 4. Economics are directionally credible
• 5. Capability implications are visible
• 6. Critical assumptions are testable
• 7. Demand and the right to win are evidenced together

Digital method, questions and traps

Include at least one non-AI alternative wherever AI is proposed, and set each option’s cost lines and economic tests from the economics of the technology model in play, as set out in Part one.

Digital method

• 1. Define option dimensions: improve the core, scale data and AI, or open new ways to compete.
• 2. Build a use-case portfolio across three horizons, anchored in value driver trees.
• 3. Screen adjacencies and platform or ecosystem plays by distance from the core.
• 4. For each option, state the platforms, data, partners and capabilities it needs.
• 5. Frame build, buy and partner routes for each major capability.
• 6. Create at least three materially different options plus a costed continuation baseline; the baseline is not counted as an option.

Advantage thesis and activity-system logic

An option is credible when distinctive activities reinforce one another around the value proposition. An advantage thesis looks like a reinforcing activity system rather than a linear chain. The illustration is Meridian option B, a premium uptime service (Meridian is a fictional teaching case, and the figures are illustrative).

The value proposition sits at the center: guaranteed availability for regulated sites. Six distinctive activities serve it:

• Failure-risk ranking from site histories
• Planned interventions before failure
• Availability-linked subscription price
• Audit-ready regulated safety records
• Shared reliability cell across sites
• Each intervention adds failure data

The activities also reinforce one another. Each intervention adds failure data, which improves risk ranking, which targets planned interventions, which makes the availability price affordable, and so on.

Illustrative. Remove any one activity and the others weaken: that interdependence, not any single activity, is the advantage. Record the thesis in Tool 9.

Test: remove one activity and ask whether the others still deliver the proposition. If they do, that activity is not part of the advantage. A useful facilitation question is which of our trade-offs a competitor would refuse to make. The common misreading is listing best practices as distinctive activities.

Comparable options and a credible baseline

Options can only be compared against an honest continuation baseline, with weights agreed before scoring. The baseline includes deterioration and obligations, not a frozen present. Agree weights and anchors before scoring starts.

Step 7 chooses a conditional strategic direction within the approved risk appetite.

Select a conditional direction and accept its trade-offs and exclusions. Record what would reverse the preference: a decision-maker should know which assumption, if wrong, changes the answer.

Step 7 · Make the strategic choices

Select a conditional strategic direction, accept its trade-offs, and document the logic and evidence that must be validated before final approval.

Workshop time: half to one day.

Completion checklist

• 1. Criteria were set in advance
• 2. The strategy addresses the crux
• 3. Where-to-play choices are bounded
• 4. How-to-win is specific
• 5. Trade-offs and exclusions are accepted
• 6. Resources will move
• 7. The choice is explicitly conditional where uncertainty remains

Digital method, questions and traps

Screen options against non-negotiable thresholds first, then weigh the survivors. A high weighted score never offsets a fatal weakness. Test how sensitive the ranking is to the weights.

Digital method

• 1. Apply thresholds: value, safety, legality, security, architecture fit and resilience.
• 2. Score survivors on fit, advantage, economics, deliverability, reversibility and evidence confidence (Tool 8).
• 3. Run scenarios and sensitivity on decisive assumptions; state what evidence would reverse the ranking.
• 4. Confirm the protected platform allocation before funding new work.
• 5. Record build, buy or partner choices and the exit route for each partner.
• 6. Take the choice, its exclusions and staged commitments to the forum holding direction authority.

Divergence before evidence-based convergence

Options are widened deliberately, then narrowed against thresholds and evidence agreed in advance. Widening the field comes first: options that differ materially, not variations on one answer. Narrowing then follows agreed rules rather than preference. Non-negotiable thresholds remove options outright, the survivors are compared on the same evidence and scenarios, and each is put under deliberate challenge. The choice that emerges is conditional, held with the tests and conditions that would confirm or overturn it.

• 1. Diverge: materially different options
• 2. Screen: non-negotiable thresholds
• 3. Compare: evidence and scenarios
• 4. Challenge: counterevidence and responses
• 5. Select: conditional direction
• 6. Validate: bounded tests and conditions

Where to play and how to win are generated as pairs

Each alternative couples an arena with a way to win and is tested against one costed baseline. Where to play and how to win are mutually dependent, so they are generated together. The illustration places Meridian’s alternatives on one grid by arena and way to win. Validation later narrowed B to six class A sites, D stayed the live comparator, and class B was paused because its failure data was too sparse.

The illustration holds baseline cash flat at $0.40m a year. A real baseline includes unavoidable reinvestment, deterioration and ongoing obligations; omitting them flatters every alternative.

Each pair records

Rules

• Generate at least two materially different pairs in addition to the baseline.
• Do not confirm an arena until its way to win has been tested.
• Compare every pair on the same scenarios, horizon and baseline.
• Keep rejected pairs and reasons in the decision record.

The grid is an executive view of Tools 6, 8 and 9, not a separate register: each cell points to its option card. Figures are illustrative.

Facilitation question: for our favored arena, what is the strongest alternative way to win, and why did we reject it? The common misreading is confirming an arena before any way to win has been tested.

Fund learning only when its result could change the decision

Each decisive assumption records what a test costs, what it can reveal and which result changes the choice. One rule connects assumptions, scenarios, pilots and conditional approvals: fund learning when better evidence could change the decision by more than the test costs in money, exposure and delay. The record is kept in Tools 7 and 17; here it is completed for Meridian’s price and cost assumption before the Direction gate.

Decisive assumption record (Tools 7 and 17): Meridian price and cost

The decision rule

• 1. Could a plausible result change the choice? If no, decide now and monitor the indicator.
• 2. If yes: is the decision cheap to reverse? If yes, decide and set a review trigger in Tool 17.
• 3. If no: is the test cheaper than the exposure it can avoid, and fast enough? If no, decide on bounded scenarios and affordable loss.
• 4. If yes: fund a capped test with thresholds agreed before results (Tool 7).

Facilitation question: which result of this test would make us change our mind? If none, why are we testing?

The strategic direction gate selects a conditional direction.

It releases only capped validation work, not full implementation. Check that exclusions, decisive assumptions and validation conditions are explicit.

Strategic direction gate

Select a conditional strategic direction and accept its exclusions, assumptions, risks, resource consequences and validation conditions. The forum is the governing body or executive decision forum, and the gate applies tenets 3, 4 and 6.

Gate checklist

• 1. The choice addresses the crux
• 2. Where and how to win are clear
• 3. Alternatives and exclusions are recorded
• 4. Resources will move
• 5. The advantage thesis is explicit
• 6. Validation conditions and expiry dates are recorded

Target capability profiles follow from the strategic choices

After Step 7, each capability the strategy depends on gets a target profile: a maturity level on one operational scale, plus its strategic importance and advantage evidence. Maturity uses one operational scale from Initial to Optimizing.

Rate maturity, importance and advantage separately

Rate each dimension on its own evidence. A capability can be mature without differentiating, or distinctive while still emerging.

Three test cases show why the ratings stay separate: a mature capability can be a commodity, an emerging one can already be distinctive, and an essential one may need high reliability without any differentiation.

Each choice traces to capabilities, architecture and transitions

Architecture implements strategy; it does not replace it. Record only the domains a choice materially changes. Walk the chain using Ardent Mutual (a fictional insurer; every figure is illustrative): strategic choice, required capability, architecture implication, transition state, investment, measure. Target maturity follows strategic need, so not every capability goes to level 5, and the architecture domains are prompts, not eight required deliverables.

Tools 12, 24 and 25 record each gap with its current, interim and target state, price, dependency and owner. Technical debt, coexistence and decommissioning obligations are costed in the transition, not left to operations.

A business model is ready to scale only when every link from customer to capital holds.

Step 8 shows how the chosen strategy creates, delivers and captures value, and tests the assumptions that decide viability before scale is authorized.

Step 8 · Design and validate the business model

Show how the chosen strategy creates, delivers and captures value, and test the assumptions that determine viability.

Workshop time: one to three days plus tests.

Completion checklist

• Customer and payer are distinguished
• The value proposition is evidenced
• Delivery components reinforce each other
• Economics reconcile
• Critical assumptions are ranked
• Tests have thresholds
• The model can be stopped or adapted at defined points

Digital method, questions and traps

Step 8 designs how the chosen digital moves create, deliver and capture value. For digital moves, the business model must answer who pays, what AI costs to run and who owns the data and models, and the unit economics must hold at scale, including consumption costs.

Digital method

• 1. Specify the customer, the job served and the digital value proposition.
• 2. Redesign the business model blocks the choice changes, using the Business Model Canvas.
• 3. Model unit economics, including run cost, cost to serve and AI inference cost.
• 4. Define platform and partner roles, and data rights, for each ecosystem move.
• 5. Test the riskiest assumptions with pilots and minimum viable products.
• 6. Write the benefits case with the business owner and Finance.

Business-model coherence

Value creation, delivery and capture must hold together before scale is authorized. A business model holds together only if each link supports the next.

• 1. Customer and payer
• 2. Value proposition
• 3. Delivery system
• 4. Revenue and cost
• 5. Cash and capital
• 6. Scale or revise

Name who the customer is and who actually pays, state the value proposition that earns that payment, and show the delivery system capable of producing it repeatedly. Revenue and cost then have to work at unit level, and cash and capital have to survive the ramp.

Scale is authorized only once those links hold; where one breaks, the model is revised rather than funded. If any link fails, the model is not ready to scale, however strong the other links are.

Building the economic bridge

Each line of the economic case needs one working definition, so claimed value survives challenge. Agreed definitions stop claimed value from evaporating under challenge from Finance.

Business value model: one value model per outcome, from baseline to verified result

Adoption is a leading indicator; only a validator can turn forecast value into realized value. Every material strategic outcome has one traceable value model, built from Step 1 and verified in Steps 11 and 12. Five parts sit above a seven-stage chain: objective and baseline; value driver and mechanism; estimated benefit; cost and time to value; ownership and verification.

The status row keeps three things apart: forecast value from the model (modeled value, scenario range and confidence), observed operational improvement, and validated realized value confirmed by the category’s validator. Adoption is a leading indicator; it never proves financial value by itself.

Adapted from the five-part value model in the Vision to Blueprint method and integrated with this playbook’s benefit tools.

Economic appraisal: the appraisal horizon follows the investment, not a default

Options competing for one decision share one horizon; unequal lives are normalized. Value that is not cash is stated, not forced into NPV. Choose the horizon from the investment’s economic life and decision context, and state risk reduction, mandatory obligations and option value separately.

Ask which assumption would reverse the preferred option: that is its switching value.

Pass, fail and inconclusive are three different results

A validation result must be classified honestly, because each of the three outcomes demands a different decision. An inconclusive result is not a pass; it calls for redesigned tests or a narrower commitment.

The operating model must deliver the business model under ordinary conditions, not only in a pilot.

Step 9 configures the organization so its capabilities and management system can repeatedly deliver the business model and strategic advantage, with decision rights assigned before the organization chart is redrawn.

Step 9 · Design the operating model

Configure the organization so its capabilities and management system can repeatedly deliver the business model and strategic advantage. The outputs include a people-and-change assessment covering roles, skills, incentives, readiness and leadership actions.

Workshop time: two to four days.

Completion checklist

• Outcomes anchor the design
• Capabilities and processes are end to end
• Decision rights are explicit
• All operating-model dimensions align
• Distinctive gaps are prioritized
• Transition states protect delivery
• The Coherence gate is passed

Digital method, questions and traps

Step 9 designs how business, digital, data and AI roles deliver the chosen business model in normal operation, not only in a supported pilot. Work through the method in order: outcomes and service levels first, then the governance pattern and decision rights, the data and AI interface, product teams and architecture authority, capability funding, and people and change.

Digital method

• 1. Define the business outcomes and service levels the operating model must deliver.
• 2. Choose a governance pattern (see the four reference patterns at the end of this step); set decision rights, blocking rights and waivers in Tool 20.
• 3. Agree the Data & AI interface charter: who decides, who builds, who reports.
• 4. Design product teams, the front door and architecture authority, central or delegated by domain.
• 5. Fund stable capabilities as products and show business units the cost they consume.
• 6. Assess people and change: who must behave differently, skills and reskilling, incentives, readiness, adoption owner and leadership actions (see the adoption stages in Step 11).
• 7. Name which business and technology changes must move together.

Separate governance (direct and control) from management (plan, run, improve), so that decisions inside delegated limits stay with management. Confirm target capability profiles here, after the conditional direction; full gap costing follows in Step 10. Assign decision rights to real roles: not every enterprise has a separate Data and AI policy owner. The common misreading is redrawing the organization chart before agreeing decision rights and the work the model must do.

The operating model has ten connected dimensions

It must deliver the chosen business model under ordinary conditions, not only in a pilot. The target operating model is ten connected dimensions rather than a technology stack. Capabilities enable the value streams; data and technology enable the capabilities. People and partners deliver them. Enterprise architecture keeps business, data, application and platform layers coherent; it is the coherence mechanism across business, information, data, applications, integration, platforms and infrastructure, not only technology design.

Illustrative content from the Meridian example. The ten dimensions apply to every strategy product. The operating states row makes the interim state explicit: Meridian runs six class A sites with routine teams before any scale-up. Facilitation question: which dimension would fail first if pilot support were withdrawn tomorrow? The common misreading is designing the technology layer and assuming the rest will follow.

Digital’s mandate has three tiers of authority

Innovation can be owned by one function. Transformation changes processes, roles and decision rights across the business, so no single function can own it; it needs shared accountability.

Digital shapes and prices demand; the portfolio decides

Decision rights from demand to architecture, shown here in a centralized functional pattern. Digital turns requests into outcome statements and prices them, but gives up any informal veto over demand.

Each function holds distinct rights from data to support

Delivery closes at business acceptance; the product team reports against the benefits case until each benefit is validated (Finance for monetary value). Benefits count only when validated in operations.

Technology risk is business risk, owned by the business

The board sets direction and risk appetite. Digital operates as the first line for the technology and cyber risks it runs and for the data and AI controls it operates. The accountable business risk owner accepts residual exposure within delegated authority and approved appetite.

A CRO approval or challenge does not transfer ownership of the risk. Where an organization formally delegates risk acceptance differently, record the delegation and keep second-line independence intact. Facilitation question: for our largest technology risk, who signs the residual risk acceptance today?

Governance outcomes are mandatory; the structure is a choice

Every enterprise needs the ten minimum governance outcomes. Which committee, function or reporting line delivers them depends on the enterprise: centralized, federated, product-based and hybrid patterns can all deliver them.

Minimum governance outcomes

• 1. Business outcome accountability
• 2. Investment prioritization and authorization
• 3. Enterprise and solution architecture authority
• 4. Data ownership and governance
• 5. AI governance and model approval
• 6. Cybersecurity and resilience decisions
• 7. Risk appetite and risk acceptance
• 8. Exception and waiver authority
• 9. Operational service ownership
• 10. Benefits ownership and independent validation

Governance directs and controls; management runs and improves

They are complementary, not interchangeable: decisions inside delegated limits stay with management. Separating the two lets executive forums decide what only they can decide.

Governance sends direction, appetite, policy and delegated limits to management. Management returns performance, conformance, risk and benefit evidence, and escalations.

Who decides: illustrative cases

Sources: ISO/IEC 38500 (evaluate, direct, monitor); ISACA COBIT 2019 (governance objectives EDM; management objectives APO, BAI, DSS, MEA). The application here is this playbook’s. Facilitation question: which decisions reach our executive forum that should have stayed with management, and which stayed below that should have come up?

Nine decision-rights terms, used one way in every pattern

Accountability for an outcome is not authority to approve a technical or regulatory decision. Tool 20 uses these terms only. The most common confusion is between owning an outcome and approving a technical or regulatory decision.

Policy authority (who sets the rule) is distinct from decision authority (who applies it to a case). The minimum governance outcomes earlier in this step are mandatory; who holds each right is configured in the reference patterns that follow.

Four reference patterns place the same rights differently

Examples to configure, not requirements. Each cell names who approves; owners, challengers and blockers are set in Tool 20. Choose the pattern that fits the organization and record it in Tool 20.

Same rights, different holders. Centralized assignments in the examples are illustrations, not requirements.

Coherence confirms the scale the strategy can bear, and releases no funding.

The Coherence gate confirms the complete strategy once business-model viability and operating-model deliverability are evidenced. Funding remains a separate decision at the Commitment gate.

Coherence gate

The executive team, with business and operating owners, confirms the strategy is viable and deliverable at the specific scale proposed, on six coherence tests. The gate applies tenets 4, 5 and 6. Check that material gaps have owners and deadlines before proceeding.

Gate checklist

• Value creation, delivery and capture cohere
• Critical assumptions meet evidence thresholds
• Capabilities and decision rights align
• Transition protects current operations
• Six coherence tests set the maximum scale
• The advantage mechanism survives both model tests

Six coherence tests set the scale the strategy can bear

The integrated strategic coherence assessment runs six tests at the scale proposed. Each test records its evidence, result, confidence, residual uncertainty, owner and corrective action. Meridian’s six-site extension is shown.

In Meridian, three results are Conditional and one Inconclusive, so the justified scale is six class A sites. The forum still decides. Ask whether the pilot relied on exceptional support.

Step 10 is where strategy becomes money and capacity.

Step 10 converts choices into an outcome logic, measures, funded portfolios, capacity allocations and accountable commitments, and proves that the selected portfolio fits both the cash and the capacity limit.

Step 10 · Translate strategy into measures and investments

Convert choices into an outcome logic, measures, funded portfolios, capacity allocations and accountable commitments.

Workshop time: one to two days.

Completion checklist

• Objectives express outcomes
• Measures include baselines and definitions
• Portfolio items trace to choices
• Funding and capacity reconcile
• Trade-offs across horizons are visible
• Benefit owners accept accountability
• The Commitment gate is passed

Digital method, questions and traps

Step 10 turns the digital choices into objectives, measures and a funded portfolio. Each measure must genuinely represent its construct.

Digital method

• 1. Link objectives to business results with a strategy map.
• 2. Classify each measure as KPI, KRI, benefit or countermetric, leading or lagging; test principle coverage across the portfolio.
• 3. Build a value driver tree from each measure to the levers teams control.
• 4. Structure the portfolio into run, change and innovate, with the protected allocation.
• 5. Fund innovation separately with stage gates and kill criteria.
• 6. Agree how Finance will confirm benefits in operations.
• 7. Pair outcome measures with flow measures such as lead time and time to restore.

Give each measure a baseline, owner, formula, source, target and response threshold. Adoption is a leading indicator, not an efficiency measure.

Execution control is a closed loop that ends in decisions

Five instruments and one evidence review form a single loop; reporting without a decision is not control. Read it clockwise: the scorecard states outcomes; OKRs turn them into period commitments; KPIs and controls run operations; the variance review routes evidence; the portfolio forum funds, defers, resequences or stops; benefits and learning test whether the case still holds.

• 1. Strategy map and scorecard
• 2. Short-cycle objectives or OKRs
• 3. Operational KPIs and controls
• 4. Evidence and variance review
• 5. Portfolio and funding decisions
• 6. Benefits and strategic learning

Every review ends in a decision: continue, adjust, reallocate, escalate or stop. Two feedback routes leave the loop. Strategic choices: a broken strategic assumption reopens Steps 6 to 9 through their gate. Resource allocation: portfolio decisions fund, defer, resequence or stop.

Who reviews what, and what it can change (Tool 14)

Tool 14 records the loop; Tool 18 runs the quarterly realization review; Tool 17 holds the triggers. Cadences are defaults to adjust to the decision. Facilitation question: which of our current reports has not changed a decision in the last two quarters? The common misreading is treating a dashboard as control.

Specify measures before using them to govern

A measure can only govern once its definition, source, baseline and owner are fixed in writing. Use this field list as the minimum standard for Tool 22.

Attractive initiatives do not add up to a feasible portfolio

Five stages separate eligibility, attractiveness and feasibility before anything is funded. They keep apart three questions that weighted scoring usually blends: is an investment eligible, is it attractive, and is the whole set feasible?

The staged sequence adapts the portfolio optimization logic of the Vision to Blueprint method to this playbook’s tools. Facilitation question: which item would we fund on score alone that cannot deliver without something we have not funded? The common misreading is treating the ranked list as the funding sequence; the illustration that follows shows why it is not.

The ranking is not the funding sequence

A lower-scoring data platform is funded first because the top-scoring portal cannot work without it. The customer portal (P2) scores highest at 81, but it depends on the data integration platform (P1), which scores 62 because it creates no benefit on its own. Funding by score would launch the portal without its data.

The selected set funds mandatory and sustain work first, then P1, a reduced-scope P3 that also uses P1 data, portal design, and a capped AI pilot kept as an option. The portal build is released in Wave 2 when P1 passes its integration test.

Reconciliation: this year 1.5 + 1.0 + 3.0 + 1.0 + 0.5 + 0.5 = $7.5m of $10.0m, and 6 + 4 + 14 + 4 + 3 + 2 = 33 of 40 specialist months. Adding P5 ($3.0m, 10 months) would need $10.5m and 43 months, breaching both limits. The $2.5m is unreleased headroom, not spending. Illustrative portfolio adapted from a Vision to Blueprint method example. Figures are invented for teaching and reconcile with the extended Tool 15 matrix that follows. Facilitation question: which of our top-ranked items depends on something unfunded?

The selected portfolio fits both the cash and the capacity limit

The extended Tool 15 records dependencies, capacity, wave and decision alongside the five categories. The “Selected this year” row totals the work selected for this year: $7.5m of a $10.0m limit and 33 of 40 specialist months. Items below that row are released later, deferred or stopped.

Fields the extended Tool 15 records

• Eligibility screen result and enabling-dependency flag
• Attractiveness score, with evidence confidence beside it
• Dependencies: the INV-nn it needs and the INV-nn it enables
• Cash, specialist months and change load per wave
• Decision: fund, defer, resequence, reduce scope, stop or option
• Funding wave and the evidence condition that releases it
• Outcome owner and the B-nn benefit it serves, counted once
• Recalculation trigger: cost, capacity or assumption change

Illustrative. Check: 1.5 + 1.0 + 3.0 + 1.0 + 0.5 + 0.5 = $7.5m and 6 + 4 + 14 + 4 + 3 + 2 = 33 months. P2 totals $3.5m and 12 months across its two waves. The $2.5m difference to the limit is unreleased headroom, not authorized spending; releasing it requires a decision within delegated limits. Facilitation question: if specialist capacity fell to 30 months, which item would we resequence first, and why? The common misreading is reading the score column as the order of funding.

The Commitment gate commits money, capacity and ownership to the chosen portfolio.

The gate commits funding, capacity, leadership attention, measures and benefits ownership. Later tranches remain conditional on evidence from the earlier ones.

Commitment gate

The executive investment forum commits funding, capacity, leadership attention, measures and benefits ownership to the chosen portfolio. The gate applies tenets 7, 8 and 9.

Gate checklist

• Every major investment traces to a choice, obligation or essential service requirement
• Funding and capacity reconcile
• Stop decisions are included
• Owners accept outcome accountability

Delivery is judged on usable capability and evidenced outcomes, not on activity completion.

Step 11 coordinates programs and projects so delivery stays aligned with strategic priorities. Funding follows evidence: teams earn the next tranche by showing usable capability and outcomes.

Step 11 · Execute the change portfolio

Coordinate programs and projects so delivery stays aligned with strategic priorities and produces usable capabilities and outcomes.

Workshop time: a recurring monthly forum.

Completion checklist

• Programs have outcome charters
• Dependencies are managed across the portfolio
• Traceability is current
• Benefits and assumptions are reviewed
• Capacity constraints are visible
• Stop and redirect decisions occur
• Operations participates before transition

Digital method, questions and traps

Step 11 delivers the digital portfolio through joint product teams and stage gates. Sequence waves around platform foundations, dependencies and risk, and plan adoption alongside build.

Digital method

• 1. Turn choices into product charters with outcomes, owners, gates and a strategy type.
• 2. Sequence waves around platform foundations, dependencies and risk.
• 3. Build to quality, security and safety gates; hold releases that fail them.
• 4. Embed AI in workflows with human oversight, limits and the ability to stop it.
• 5. Track adoption, benefits and cost in short cycles; stop what does not prove value.
• 6. Route new ideas back through the front door.
• 7. Set retirement dates for the legacy processes and platforms each wave replaces.

The strategy-to-value system

Delivery is judged on usable capability and evidenced outcomes, not on activity completion. Owners and evidence connect every link; necessary obligations have an explicit justification path.

• 1. Choices and objectives
• 2. Required capabilities
• 3. Funded investments
• 4. Usable delivery increments
• 5. Operational adoption
• 6. Realized outcomes

Each link is only credible if the one before it holds. Choices and objectives define what is worth funding; required capabilities say what the organization must be able to do; funded investments commit the money to build them.

Usable delivery increments are the first point at which anything can be used, operational adoption shows the business has taken it up, and realized outcomes are the evidence that value arrived. A break in any link stops value at that point, whatever the plan reports.

Deployment is not adoption, and adoption is not benefit

For each material change, name who must behave differently, what makes that possible, who owns adoption and what evidence shows it lasts once pilot support is withdrawn. Seven stages run from deployment to realized benefit.

Step 9 sets role impacts, skills, incentives and workflow redesign (Tool 12). Step 11 manages readiness, resistance and adoption risk. Step 12 withdraws exceptional pilot support before sustainment is judged and records the stage reached in Tool 16. Logins and training attendance show availability or use, not adoption.

Each benefit category has its own validator

Finance keeps authority over every monetary claim. Nonfinancial value is real when its accountable owner validates it against an explicit evidence standard.

Benefit states: forecast, modeled, observed, validated, realized, sustained. Only validated benefits are reported as realized; sustained means evidenced again at 6 and 12 months. A shared-platform benefit is attributed once, across all initiatives that claim it. Mandatory ethical, safety and regulatory requirements are thresholds, never offset by a weighted score.

The portfolio stops work to fund better work

Run, change and innovate describe the work; the five investment categories decide how it is funded and when it stops.

Each quarter (Tool 18) the forum reviews intake, traceability, capacity, affordability, risk, dependencies and timing, and records a continue, stop or reallocate decision for every material item in Tool 15. Status reporting is an input to that review, not its agenda. The category rules are as set out in the five portfolio categories in the annexures. The quarterly review is a decision meeting, not a status report.

Responding between scheduled reviews

Triggers and response authority are agreed when the strategy is approved, not after a threshold is breached. That way the organization can act between scheduled reviews without improvising authority.

Adaptation reopens only what the evidence invalidates.

Step 12 embeds new capabilities in operations, tests whether outcomes materialize, and keeps the strategy current. Benefits cannot be claimed because a project or pilot is complete.

Step 12 · Transition, learn and adapt

Embed new capabilities in operations, test whether outcomes materialize, and adapt the strategy when evidence invalidates assumptions.

Workshop time: quarterly and on trigger.

Completion checklist

• Acceptance criteria are met
• Operational ownership is explicit
• Adoption and performance are stable
• Benefits are evidenced
• Assumptions and scenarios are reviewed
• An explicit continue, scale, adapt, pause or exit decision is made
• Learning updates the strategy system

Digital method, questions and traps

Step 12 moves digital change into reliable operation and keeps the strategy current through quarterly reviews and triggers.

Digital method

• 1. Agree operational acceptance, service levels and support before final release.
• 2. Transfer run, security operations and lifecycle management to Digital support.
• 3. Test critical services against disruption tolerances, including third parties.
• 4. Validate each benefit with its category validator (see the benefit validators in Step 11) and close the case only after sustainment evidence.
• 5. Review risks through the three lines of defense and report to the board.
• 6. Refresh the strategy when signals, regulation or technology change.
• 7. Review the alignment map every quarter in the portfolio cadence.

Adaptation reopens only what the evidence invalidates

Evidence returns to diagnosis, choices or allocation; learning that reallocates nothing is not adaptation. Strategy is a governed learning loop that can reopen earlier decisions, at the right level. Read the loop clockwise from operational outcomes.

• 1. Operational outcomes and adoption
• 2. Evidence and assumption monitoring (Tool 17)
• 3. Review: periodic or on a trigger (Tool 18)
• 4. Decide: continue, scale, adapt, pause or exit
• 5. Reallocate resources or reopen a step

At the center sits value realized in normal operations. Every review ends in continue, scale, adapt, pause or exit, and the result reallocates resources or reopens the step that made the failed assumption.

Who decides, and what a trigger reopens

Control, safety and continuity breaches are acted on at once by the designated operational authority under existing procedures, then reviewed. Facilitation question: which of our assumptions, if wrong, would reopen the Direction gate, and what signal would tell us first? The common misreading is treating every change as a reason to restart the strategy, or treating no change as success.

Meridian restores growth by selling reliability, not by automating dispatch.

Part four works three fictional cases. Meridian Industrial Services is business-led and completes every step; Halden Cold Chain is digital-led; Ardent Mutual shows a technology renewal. Meridian Industrial Services, Halden Cold Chain and Ardent Mutual are fictional businesses. Every number, observation and authority limit is illustrative. The cases demonstrate reasoning and completed decisions, not a market forecast.

The mandate and the initial crux

The sponsor asks how to restore profitable growth across an initial cohort of 20 customer sites while preserving regulated-service reliability. Meridian is a fictional industrial service business; every figure is illustrative. Note that the opening hypothesis framed the work as a cost and technology problem.

The Meridian records, step by step

Each record that follows is a completed tool or decision from the method, in the order the work produced it. Open a record to read it in full.

Halden’s data team proposed a new service, not a new system.

Halden is the digital-led example. Digital and Data & AI leaders propose; the business owns the value question and decides. Halden Cold Chain is fictional. Every figure in this example is illustrative, not case evidence.

The intake record

Step 1 records a digitally originated opportunity. Digital and Data & AI leaders propose and evidence it; the commercial director owns the value question and the executive forum decides.

Evidence moved the advantage from prediction to prevention

Outside-in and inside-out evidence, reconciled at the checkpoint after Step 5. Prediction alone is widely available; prediction joined to intervention is not. That combination is what competitors could not easily copy, which shows advantage built from capabilities rather than unique data.

The platform fails at three years; five needs carrier data

USD millions, incremental to continuing today’s service; three years at 10 percent, matching the service’s contract and model-refresh life; investment in year 0. O1 shows that using AI internally does not pay by itself. The assurance service O2 leads; the platform O3 only pays over five years with optimistic growth, so it is kept as a triggered option.

Halden Cold Chain is fictional. Every figure in this example is illustrative, not case evidence. Continuation baseline: today’s service with $1.2m a year of excursion claims.

The service needs an operating model that can act, not just alert

Steps 8 and 9: the business model is tested as a service, and the agentic intervention runs at autonomy level 2 with hard limits (Tool 23), passing the AI lifecycle stages before each tranche. The limits are a $2,000 spend limit, dispatcher approval for every order and a stop rule if cost per intervention exceeds $45.

Validation narrowed the opportunity before money moved

A 12-week paid pilot on three pharmaceutical lanes changed the price and cost assumptions. The full service no longer paid, but a narrowed scope did. The business forum, not the data team, made the call and released only a conditional tranche.

Illustrative. Revised increments, USD m: O2 full 0.03, 0.32, 0.36; O2 narrowed 0.12, 0.32, 0.34; O4 0.08, 0.18, 0.17. Monetary benefits count only when Finance confirms them.

Ardent’s platform problem was change speed, not age.

Ardent is the technology-renewal example. Coupled product rules slow every change and concentrate resilience risk on one platform, so the renewal is bought in tranches and governed differently. Ardent Mutual is a fictional insurer. Every figure in this example is illustrative.

The intake record and the revised crux

Steps 1 to 5: functional Digital/IT strategy, business-led, Technology Transformation.

Progressive replacement wins only if it is bought in tranches

USD m vs the baseline, five years at 10%, the migration and contract period; seven years is a sensitivity. Downside: 25% overrun, benefits a year late. R3’s downside NPV is negative, so the commitment is limited to a $3.0m first tranche with release conditions, and a reversal path back to re-hosting if overruns exceed 25 percent.

Ardent Mutual is fictional; figures illustrative. Net cash for R2 and R3 includes later tranches: R2 spends $6.0m in year 1; R3 spends $3.0m in each of years 1 and 2 against gross benefits of 0.8, 2.0, 3.5, 4.2 and 4.8. Downside applies the 25% overrun to every tranche and delays gross benefits one year. R2 turns positive only after year 5 (seven-year NPV 4.36, against 7.83 for R3). Terminal value excluded.

Renewal changes how teams, risk and money are governed

Steps 9 to 12 for tranche 1: operating model, decision rights, measures and the rules for counting value. Ardent uses a federated pattern: the COO owns outcomes, the CIO executes, enterprise architecture can block on standards and the COO accepts residual migration risk, with the CRO challenging independently.

Twenty-five worksheets carry the minimum evidence and decisions through the method.

Treat them as working records and decision controls linked by one identifier chain, not forms to complete mechanically. Thirteen have fillable templates later in this part; all 25 are in the companion toolkit. Each card states the tool’s purpose and its owner.

Tools 1 to 6: mandate, diagnosis and options

Tools 1 to 6 bound the decision, establish the diagnosis and generate comparable options.

Tools 7 to 12: choice, models and coherence

Tools 7 to 12 test assumptions, compare options and prove the models hold together. Tool 8 screens thresholds first and records what would reverse the preference.

Tools 13 to 18: commitment, delivery and realization

Tools 13 to 18 record commitment, govern delivery and confirm value that is actually realized. Tool 16 is where a modeled benefit becomes a confirmed one, and only Finance can confirm it.

Tools 19 to 24: digital strategy

Tools 19 to 24 set up the front door, decision rights, the Data & AI interface and Digital’s measures. Tool 20 records decision rights in the chosen governance pattern; Tool 23 ties AI use cases to value, cost per task and stop conditions.

Tool 25: alignment map and gap register

Tool 25 tests fit across the four Strategic Alignment Model domains and turns each break into an owned, measured fix: a gap with a KPI and a fix.

Record the current and target state in each domain, then test the four links: Business ↔ Organization, Business ↔ Digital Strategy, Digital Strategy ↔ Digital Infrastructure, Organization ↔ Digital Infrastructure. Name a Data & AI line in both digital domains.

Gap register

Owner: Strategy lead with the Enterprise architecture lead. Used in Steps 3, 5, 9 and 12 and at the coherence gate.

Source: Henderson, J. C. and Venkatraman, N. (1993), IBM Systems Journal, 32(1). See the Strategic Alignment Model in Annex D.

Thirteen tools come with a template, and one identifier chain links them all.

Each template lists the fields to record. Where a template is pre-filled with Meridian entries, those rows are kept as worked examples. All 25 tools are in the companion toolkit.