01
Pre-specified obligations
Escalation conditions, evidence requirements, and gate criteria are defined per domain before the initiative starts. Governance triggers itself. Removes the failure of the right question never asked.
How I work
Value Governance Architecture designs the structure. AI Value Orchestration runs the AI initiative through it.
Two disciplines. One purpose: ensure AI investments remain aligned to the intended outcomes set by the board.
The context
Build is the smaller half of the problem.
Most enterprises spend on the technology layer and almost nothing on the layer that decides whether the investment converts. The research is consistent. The failure mode is coordination and accountability, not the model.
36%
of CFOs feel confident in their ability to deliver real enterprise impact from AI, even as budgets accelerate
Gartner, 2026
39%
of executives can attribute any EBIT impact to agentic AI, against an anticipated 171% return
PagerDuty / McKinsey, 2025
60%
of AI projects unsupported by AI-ready data are expected to be abandoned through 2026
Gartner, 2025
95%
of enterprise AI pilots produce little or no measurable P&L impact
MIT NANDA, 2026
51
successful enterprise AI deployments analysed
The differentiator was the organisation, its readiness, its processes, and its governance, not the AI model. Value Governance Architecture designs that readiness. AI Value Orchestration runs the AI initiative through it.
Stanford University, 2026
How I work
Eight operating principles governing how I work.
Derived from twenty years of programme accountability inside organisations and across them. Each carries the outcome it produces and the failure it removes.
VGA
Value Governance Architecture
The design discipline. It builds the accountability structure before the AI initiative runs through it.
02
Live decision rights
Every domain interface names decision-maker, evidence, and authority. Accountability is operable across domains throughout the initiative. Eliminates the informal decisions that produce most drift.
03
Portfolio-ready design
The structure coordinates domains across more than one AI initiative from the start. Scales without rebuild. Protects each investment from being cannibalised by the next.
04
Cadence, not crisis
Domain coordination is reviewed on a fixed schedule, not reopened when something changes. Governance stays stable. Becomes a standing institutional capability, not a per-initiative overhead.
VO
AI Value Orchestration
The run discipline. It moves the AI initiative through that structure in real time, on the value owner's behalf.
01
One outcome line
Every domain and workstream connects to the specific revenue, cost, or margin figure the investment was funded to move. No line, no work. Prevents investment drift.
02
Trajectory, not plan
When specifications or demand shift, the outcome line is reassessed against what is now feasible before the next board review. Gives the sponsor a defensible position.
03
Board-trusted ownership
The value owner is assigned pre-start, holds board trust, and remains the conduit. The orchestrator runs on their behalf. Ends the everyone-and-no-one accountability pattern.
04
Capped coordination
Three active initiatives is the ceiling. A fourth queues at capital allocation, not at the orchestrator's discretion. Coordination stays meaningful. Protects quality at portfolio scale.
The investment-to-outcome lifecycle
Five stages across the lifecycle.
The discipline runs through the investment-to-outcome lifecyle: from the conviction to invest through to the operation that carries the system after handover. The lifecycle is linear overall, with stages running in parallel and iterating where the work demands it.
1
Stage 1
Investment decision
Capital case validated against what prior initiatives actually delivered. Benefits baseline set. Value owner named. Every domain the initiative will touch identified before commitment. Everything downstream measures against this.
2
Stage 2
Design and build
Value Governance Architecture designed before any model is trained. Decision rights, gate criteria, and escalation paths agreed across every domain the initiative touches. Build runs inside that structure, not alongside it.
3
Stage 3
Change and adoption
Workforce, process, and adoption work runs in parallel with build across every domain. The value case assumes people work differently. Each domain is ready to operate the system, not just receive it.
4
Stage 4
Handover to live service
Accountability transferred to the named value owner with the value architecture map, benefits baseline, and domain operating structure in hand. Orchestration is temporary custody across domains. It exits when the structure holds.
5
Stage 5
Steady state
Delivered value reconciled against the baseline across every domain for twelve to twenty-four months. Value realisation lags go-live by design. Produces the record the next investment case is validated against.
Inside the lifecycle
Two disciplines run that lifecycle. Here is where each one works.
Blue is the structure. Teal is the motion through it. Amber is the board accountability surface, produced on a fixed cadence throughout, not at the end. Governance maturity shifts as the engagement runs, with the inheritance package handed to the client at close.
Stage 1
Investment decision
Stage 2
Design and build
Stage 3
Change and adoption
Stage 4
Handover to live service
Stage 5
Steady state
VGA
Value Governance Architecture
Designed once, up front. The structure governs stages 1 to 4.
Domain mapping
Decision rights allocation
Escalation design
Gate specification
Portfolio layer
VO
AI Value Orchestration
Runs the initiative through the structure, end to end, then exits.
Mandate scoping
Programme activation
Live coordination
Gate management
Readiness and handover
Post-mortem
Next step
This is the method. What I Do is what you commission against it.
Three engagement shapes, three commercial models, laid out on the What I Do page.
Perspectives on these disciplines publish via The Transformation Generalist. The academic argument is in progress at Edinburgh Business School.