The question: do you understand the P&L of this change early enough to commit to it?
Transformations fail when the financial impact lands too late. By then the money is committed, the business case has eroded, and the CFO was never really bought in. This lens puts the P&L of change on the table at the start.
Most AI cases fail not because the model is weak, but because the operating model never changes. Spreading licences widely creates activity, not value. Copilot everywhere is not a business case.
Value appears when AI is attached to a proposition or journey with measurable outcomes: roles redesigned, work reallocated, controls rebuilt, and leaders willing to move budget, power and accountability. If headcount, decision rights, controls and the service model do not change, the productivity gain stays trapped as local convenience. That is why we work from journeys and propositions, not tool rollout.
Most of the cost base is already committed. Cost the change against the addressable slice, not the whole base.
Illustrative. The as-is base spans people, process, technology, infrastructure and property, plus suppliers and committed contracts. Only the addressable part can pay for change.
When delivery was people, the doers were the cost. AI-amplified delivery breaks that ratio.
Illustrative. AI is now a cost line, to build and to run, alongside setup (cloud, pipelines, data environments) and the governance the model demands.
Investment and dual-running come first; payback comes later. Knowing where the trough sits is the point.
Illustrative. The investment-before-return curve. Transition costs include dual-running, decommissioning and contract exit; value realisation is tracked projected against actual.
The question is not "where can we deploy AI?" but "which journeys and propositions deserve scarce change capacity first?" Prioritisation is a financial discipline, not an afterthought.
We weigh each candidate on proposition value, customer or institutional urgency, risk and control reduction, and architectural enablement, against effort and transition cost. That forces trade-offs instead of letting every executive area sponsor its own AI experiment, so scarce investment goes where the return and the readiness actually are.
Financial Intelligence is not a silo. The other Intelligences are its P&L contributors; this is where cost and value meet.
Squad and supporting-staff cost, and the AI-amplified role mix.
Infrastructure run-cost, setup, and the cost of dual-running with the core.
AI build and run, and the cost of AI governance.
Data and risk governance, and the cost of control.
The benefit side: deposits, margin and cost-to-serve.
Conversion, retention and avoidable support cost: a journey can complete technically and still leak value if people misread it or never trust it. Plus the harm avoided.
Governed human-machine action made scalable, cutting the hidden supervision, control-duplication and approval-latency costs that quietly destroy AI returns.
DTA does not do your full financial impact analysis; that stays with the CFO and finance. We frame it early: the shape of the as-is, target and transition, the cost model of the operating model we design, a value hypothesis per proposition, and the sequencing, so you know when to invest, where the curve troughs, and what to protect. Enough to get the CFO bought in and to scope and time the change.
A frame and a direction, not a forecast. The audited numbers stay with you.
Tell us the proposition, architecture decision, or transformation problem you are facing. We will tell you, honestly, whether and how we can help.
Discuss a challengeField notes on agentic engineering, on-device AI and transformation. No spam, unsubscribe anytime.