The Doctrine Card

Everything this programme rests on, on one page. Print it and keep it in the folder you take to the steering meeting. Nothing here is a summary of a tool; it is a set of tests you apply to any AI proposal, in any sector, at any budget.

The six frameworks

#	Framework	The executive test
1	The Operating-Intelligence Loop sense → decide → act → learn	At which stage does our loop break? Value appears only when the loop closes.
2	The Data-Foundation Precondition	Garbage in, confident garbage out. Which single decision must this foundation serve first?
3	The Autonomy Ladder inform → advise → act → bounded autonomy	What does it touch, what happens when it is wrong, who approves, what is the undo?
4	The Value–Risk Map	Not “can we?” but “what does it cost us when it fails, and is the value worth that exposure?”
5	The Sequencing Discipline	Data foundation, bounded pilot, human in the loop, scale only what paid. In that order.
6	The Operations-Excellence Lens	Connect the capability to the result, or do not build it.

The twelve pillars, grouped by precedence

The four foundations — earn these first	The high-value pillars — only once the foundations hold
1. Executive AI fluency 2. AI-ready data foundation 3. Operational visibility Governance — named as a foundation in the doctrine; it runs through pillars 9 and 10 rather than sitting as a numbered pillar of its own	4. Predictive reliability 5. Optimisation of the core operating decision 6. Living asset or process models (twins) 7. Commercial and bid intelligence 8. Security and loss prevention 9. Audit-grade compliance and reporting 10. Cybersecurity and operational trust 11. Capability and workforce reskilling 12. Partnerships and ecosystem

They are not a menu. Attempt the exciting pillars without the foundations and the programme stalls, leaks money, or creates risk faster than value.

The prioritisation engine

Criterion	Weight	Scored 1–5 on
Value	0.35	What the loss or gain is worth, in your own units
Readiness	0.20	Whether the data and the workflow actually exist
Speed	0.20	How quickly a measured result can be produced
Fit	0.15	How close it sits to the strategy you are already funding
Risk (inverted)	0.10	(6 − risk score) — the damage when it is wrong

Priority score = (Value×0.35) + (Readiness×0.20) + (Speed×0.20) + (Fit×0.15) + ((6 − Risk)×0.10). Rank, then apply the Rule of Three: never more than three initiatives in active pilot at once.

The ninety days, and the gates

Phase	What happens	Deliverable
Days 1–30 Diagnose	An honest read: where value leaks, what data exists, what cannot be trusted, what pays fastest, what must be fixed first	AI Readiness Diagnostic — findings reports with the value at stake
Days 31–60 Design	Governance committee, data ownership, ranked portfolio, controls, budget, metrics, timeline, programme office owned by the business	The AI operating model, and a ranked portfolio
Days 61–90 Launch	Three to five pilots, each with a business owner, a data owner, a baseline and a number	Measured outcomes — and a scale-or-stop decision

Stage gates: Gate 0 deploy · Gate 1 findings · Gate 2 pilot go · Gate 3 scale go. Nothing scales on a promise.

The board scoreboard

Theme	What the board tracks
Reliability and operations	Downtime reduction · process or maintenance compliance · availability · mean time between failures and to repair · throughput or deferment loss · report timeliness · delay reduction
Commercial	Bid or proposal quality · win-rate improvement · supplier or contractor performance · cost saved · revenue protected or recovered
Compliance and data	Submission accuracy · data-quality score · issue or observation closure rate
Capability	Adoption rate · training completion

Every metric carries a baseline, an owner and a cadence. Activity metrics — models deployed, dashboards built, licences bought — are not value.

The five failure modes

Starting with technology, not value. Buy a platform, hire data scientists, change nothing in the operation.
Waiting for perfect data. Boil the ocean instead of building around the decisions that matter.
Insight without an owner. A model detects a pattern, nobody acts, nothing improves.
Automating past your ability to govern. Autonomy granted faster than control — industrialised risk.
Measuring activity instead of outcomes. The standard disguise of a failing programme.

The twenty sessions and the charter they build

#	Session	Charter section
1	Why AI Is Not Another IT Project	Section 1 — Where we are today: the honest classification and the one decision that matters most.
2	The Operating-Intelligence Loop: Sense, Decide, Act, Learn	Section 2 — Our broken loop: the stage that fails, and what it costs.
3	Garbage In, Confident Garbage Out: The Data Foundation	Section 3 — The one data contract we will sign first.
4	The Autonomy Ladder: Inform, Advise, Act, Bounded Autonomy	Section 4 — Our autonomy ladder: what each system is allowed to touch.
5	The Value–Risk Map: Fund What Pays, Govern What Bites	Section 5 — Our value–risk map and the thing we will deliberately not do.
6	The Sequencing Discipline: The Order That Survives	Section 6 — Our sequence: what we do first, second and not yet.
7	The Operations-Excellence Lens: Connect It to the Result or Do Not Build It	Section 7 — Our unit of value and the baseline we will measure against.
8	From Reporting to Visibility: Running the Business on Live Signals	Section 8 — Our operating picture and the exceptions that will interrupt us.
9	Who Owns AI: The Programme Office, Decision Rights and the Business	Section 9 — Who owns this: sponsor, office, committee, data owners, decision rights.
10	The Twelve Pillars, Sequenced Into One Roadmap	Section 10 — Our twelve-pillar scorecard and the three foundations we will fix first.
11	Governance Is Value Protection, Not Paperwork	Section 11 — Our governance standard: five controls, five signatures.
12	The Compliance Dividend: Numbers You Can Defend	Section 12 — One reported number, traced: the gaps we must close.
13	Cybersecurity, the Operational Boundary and the Shadow-AI Problem	Section 13 — Our safe-use standard and the boundary we will not cross.
14	Agentic AI and Controlled Autonomy: The 2027 Question	Section 14 — Where we will allow agency, and the line we will never cross.
15	The Board Scoreboard: Sixteen Metrics Across Four Themes	Section 15 — Our board scoreboard: metric, baseline, owner, cadence.
16	Days 1–30: The Diagnostic That Does Not Flatter Anyone	Section 16 — Our diagnostic findings and the value at stake.
17	Days 31–60: Designing the Operating Model and Choosing What to Fund	Section 17 — Our ranked portfolio and the first three we will run.
18	Days 61–90: Pilots That Carry a Number	Section 18 — Our first pilot charter, signed.
19	Capability, Partners and the Five Failure Modes	Section 19 — Our capability plan, partner test and pre-mortem.
20	The Leadership Mandate and Your 90-Day Charter	The whole charter, assembled and signed.

ERW CommandAI · built on The Intelligent Barrel — An Operating Doctrine for AI by Ritchie Wingo. Back to the programme home →

THE AI LEADERSHIP PROGRAMME

AI is not another IT project. It is your operating model.