Transformation management
The craft of running a transformation that actually transforms.
Most transformations do not transform. They reorganise, they rebrand, they rebadge the roadmap, and eighteen months later the P&L looks much like it did before, minus the consulting fees.
The craft of running one that actually transforms is not glamorous. It is a small set of unfashionable disciplines: a named sponsor who stays in the seat, a defined end state that is specific enough to be falsified, a short list of decisions that get made rather than deferred, and a willingness to retire things, not just launch things.
AI makes some parts of transformation faster — analysis, communication, documentation, training content. It does not make the hard parts easier. The hard parts are political, and politics do not scale with tokens.
The failure mode we see most often is confusing the AI tooling with the transformation itself. The programme ships impressive dashboards and dies in the operating model. The organisation chart does not change, the incentives do not change, the committee structure does not change, and the AI quietly becomes a layer of productivity theatre over the old shape.
A transformation that transforms leaves the organisation looking different to the people inside it. If the staff a year in cannot describe what changed about their work, nothing changed.
The decision: define the transformation by what the organisation will no longer do, not by what it will add.