The technology changes every decade; the principles do not. Five principles that decide whether a transformation succeeds, the watchouts we see in practice, and what each one means now that AI is the agenda.
The principles behind successful digital transformation are well understood, and they have barely moved in twenty years, while the technology has changed completely. The hard parts are recognising the complexity hidden behind ideas that sound simple, and assembling the different actions into one coherent, compelling programme. The failure rate shows how hard: research from McKinsey and BCG has put the share of transformations falling short of their objectives at around 70% for years, and BCG’s analysis of 850 companies found only around a third fully meeting their targets. In 2026, with AI now the transformation agenda for most organisations, the principles matter more, not less. Here are the five we hold to, with the watchouts we see in practice.
Principle 1: there is no digital strategy, just strategy
Digital thinking touches every element of a business: it can modify the business model, create products, reshape the organisation, sharpen operations and change the customer experience. All of that is simply strategy, and treating “digital” as a separate document guarantees it stays at the margins. The same is now true of AI, only more so. An AI strategy bought as a slide deck is doubly hollow; the useful question is what the business is trying to achieve, and where digital capability, AI included, changes what is possible.
The watchout: not every business has a well-defined strategy to plug into, or experienced people free to develop one. In those cases the transformation has a broader scope than anyone expected, because the strategy work has to happen first. We have seen this be hardest in businesses that are founder-led or that have grown by acquisition, where the strategy has lived in someone’s head.
Principle 2: internal managers need to lead the programme
The biggest hurdle in any change is understanding what is actually required, and actual processes have a habit of differing from documented ones. Internal managers can get to the truth. Change also needs advocates, and it needs to stick after any external help leaves; both point the same way. The evidence agrees on where the effort belongs: organisations that invest properly in the people and culture side of transformation succeed at several times the rate of those that treat it as a technology exercise. AI adoption follows the same law: tools that teams choose and champion get used, tools that are imposed get ignored.
The watchout: the true cost of internal leadership is under-budgeted, routinely. The managers who should lead are already at full capacity, so leading the change means backfilling their day jobs, giving them a real voice early, and investing in their capability. When resourcing is addressed late, the programme stalls and the blame lands on the technology.
Principle 3: the transformation story needs quantified goals
A transformation may span business model changes, process redesign, new customer experiences and new ways of working, all at once. Without a simple narrative and clear numbers, that breadth becomes confusion. The goals need to be quantified, cascade into personal objectives, and be understood as vital to the organisation’s future. This discipline matters double for AI, where enthusiasm produces pilots by the dozen: if the time an assistant saves never shows up anywhere measurable, the programme is theatre.
The watchout: it is always easier to describe the vision than to commit to outcome numbers, and iterative work makes targets feel premature. Set them anyway. Measurable targets make communication simpler, collaboration easier and governance honest, and yes, they make failure visible. That is what they are for.
Principle 4: leaders need to get it, and commit to it
Understanding what digital and AI mean for the business model, the proposition and the market cannot be delegated. The gap is visible in the adoption numbers: the British Chambers of Commerce reports over half of UK SMEs now using AI, while tighter official definitions put genuine adoption far lower. The distance between those numbers is mostly a literacy gap at the top: tools in use, without leadership that understands them well enough to direct them.
It can be closed quickly. In our AI literacy work we have watched leadership teams move, in a handful of focused sessions, from wary to able: able to see where AI genuinely helps their business, able to set guardrails, and able to lead the programme rather than defer to it. The prerequisite is the leader showing up to learn.
The watchout: some executives do not feel the need to develop digital competency, or fear being seen struggling with it, and many successful leaders built their careers on setting direction rather than empowering teams, which is not how transformation works. The block is human, not technical, and a good coach or the right technology leader can clear it.
Principle 5: transformation is the management of the unknown
You cannot specify the end state at the start; that is what makes it transformation. Working solutions, tested in iterations by empowered cross-functional teams, generate the data and insight that shape the next step. The tension between business targets and iterative flexibility is resolved by governance: clear expectations, honest metrics, frequent transparent communication, and prioritisation that stays aligned to outcomes.
The watchout has a new face. The old version was agile flexibility trampling governance. The new version is shadow AI: your team is already using AI tools, whether or not the organisation has decided anything. Pretending otherwise is not governance. The answer is the same as it ever was: clear guardrails, visible priorities, and leadership that channels the energy rather than banning it.
The recurring lesson
The recurring theme of failed transformation, from the business process re-engineering of the 1990s through CRM, ERP and digital, is that programmes which deprioritise change management fail. AI will not be the exception; it will be the most emphatic demonstration yet. The narrative, the ownership, the numbers, the leadership and the governance are the programme. The technology, this decade as every decade, is the easy part.
