Category: Strategy & Advisory

Severed stone aqueduct spans reaching across an empty gap, structural break at the central load-bearing joint

Plan-Build-Run is why your AI program keeps producing pilots and not products

The pilot-not-product pattern in enterprise AI is generated by the Plan-Build-Run operating model. One question changes what a CIO is actually approving.

Two identical lattices, one bronze one steel: org structure and system architecture are the same shape.

Conway’s Law is running your AI architecture whether you’re using it deliberately or not

Conway’s Law says systems mirror the org chart that built them. For AI, that means the org design you sign off now decides the architecture your teams can produce.

A precision-machined steel structural bracket rests on a matte concrete plinth in a dark studio, its finished polymer outer wrap peeled back mid-application to reveal the load-bearing steel form and bolted joints underneath, with a single bronze-toned tension bolt catching directional light against a midnight-black background.

Product-centric delivery is a funding decision before it is anything else

Gartner says 85% of organisations have declared product-centric delivery. Most have changed the language, not the model. Three structural decisions determine whether the shift has actually happened.

A precision steel dial instrument with a bronze calibration ring rests on a worn dark oak workbench beside an open ledger.

Your people are working around the AI you bought

AI adoption stalls on unchanged workflows, incentives and trust conditions — not the technology. What executives must redesign to unlock enterprise value.

Overhead view of a large partially-poured concrete substructure, arranged as a precise grid of load-bearing footings. A single warm inspection lamp casts a bronze pool of light across the centre of the grid, illuminating a folded architectural drawing resting on the formwork. Everything else is deep midnight blue and cool graphite. No people are visible.

When the AI hasn’t delivered, implementation is the second question

When an AI program has not delivered, the reflex diagnosis stops at implementation. The load-bearing question sits one layer up — at AI investment approval, where the specification the committee approved either enabled a defensible review or foreclosed one.

Dark industrial relay with single bronze-lit checkpoint among many idle gates, depicting coordination overhead at scale.

Your next technology dollar will mostly fund coordination

Scaling a technical IT operating model multiplies handoffs between specialised functions, so most new investment is absorbed by coordination before it reaches delivery. The executive who names that as a model question, not a resourcing question, has a different set of options before the next budget cycle lands.

Funding AI on a project budget is funding the appearance of progress

Enterprise AI programs fail because CIOs treat AI as a project. Run it as a product instead – with persistent teams, outcome-based funding, and governance built for continuous adaptation.

A precision brass calliper open on dark slate, measuring a hairline gap — an instrument that reads what the eye cannot see.

When your AI transformation ends and you’ve still got no outcomes

An independent review commissioned after a transformation program rarely names execution failure. It names the constraint misidentified at the start.

Cold concrete and steel governance scaffold meeting a finer machine-paced lattice along a single bronze vertical seam.

The governance structures you built were designed for slower AI

Most AI accountability structures were designed for a pace of decisions that AI has already moved past. The board-ready answer is a structural design check, not another policy refresh.

Steel scaffold engulfed by streams of bronze particles representing AI consumption overwhelming plan/build/run IT funding.

Your AI run-rate does not match the plan because the funding model assumes assets

Token-based AI spend is opex, per-call, and demand-driven — structurally incompatible with plan/build/run. The CIOs who put a service catalogue, chargeback, product funding, and AI FinOps in place this year take a defensible position into the next budget cycle.