There are three persistent misconceptions about culture.

The first is that strategy and culture can be developed independently — hence the well-known line that culture eats strategy for breakfast.

The second is that culture only changes when you replace people, as if organisations are inherently either this or that, and cannot be enriched by their opposite.

The third is that culture is too vague to manage.

We take a different view. Culture is the context within which people in an organisation interact — an enabler for every business relationship, and a minefield for every conflict. When the hard side of business and strategy clashes with culture, culture always wins. Which means culture is not too vague to manage. It is too important not to.

When culture fails, it fails the same way: a value gets cut off from the value that balances it. Speed without safety becomes recklessness. Safety without speed becomes bureaucracy. Consensus without decisiveness becomes paralysis. Decisiveness without consensus becomes autocracy. 

As Trompenaars put it: "Any value disconnected from its opposite leads to a pathology." The pathology is an organisational design problem — one that no amount of leadership training, engagement survey action, or values refresh will fix until the opposite value is brought back into the room.

AI makes this problem more urgent: organisations are being pushed to move faster, experiment more, automate more, and redesign work around digital colleagues. But if speed is not reconciled with safety, experimentation with discipline, and automation with human judgement, AI does not improve execution. It accelerates the existing pathology

Take grid operators — building a physical power plant is a different challenge from developing the digital solutions and new collaborations required for the energy transition. Or take a sector facing disruption, where a clear tension emerges between short-term and long-term, as codified in Clayton Christensen's The Innovator's Dilemma. Hold on to one orientation without enriching it with its opposite, and it becomes a weakness — or even leads to failure.

AI adds a new layer to this. It changes the art of the possible. Products, services and internal tools that once took months to build can now be prototyped in days. Processes that once required direct human execution may increasingly be handled by agents, with humans involved where judgement, trust, safety or compliance still matter. That forces a different set of strategic questions: where do we need to move faster, where do we need stronger control, where should human judgement remain decisive, and where could digital colleagues take over meaningful parts of the work?

Work on culture therefore does not start with a values session or generic leadership principles. It starts with strategy. Every conversation about culture, and every initiative to shape it, begins with the business issue at stake. Not with a values exercise. Not with a leadership programme. With the question the strategy is forcing on the organisation. We open with three:

From the answers, the dilemmas that matter come into focus. One is always in the room: the tension between the culture you have today (the "ist") and the culture your strategy now needs (the "soll"). In AI-enabled transformation, that tension often becomes sharper: the organisation may need more experimentation, faster decisions and greater openness to learning, while also needing stronger safeguards, clearer accountability and more trust in how work is redesigned.

We do not treat that as a diagnosis to produce a deck. We treat it as a reconciliation to work through — because culture has to support the business, not the other way round.

Participants work with one value at a time. They agree what the value means in their own context, then define the concrete, observable behaviours they want to see from each other — and the behaviours they clearly do not. The group prioritises, debates, and agrees how they will hold one another accountable.

Take a value like entrepreneurship through flawless execution. A team using V2B might agree that entrepreneurship means bringing a half-formed idea to the table without needing a deck — and that execution means owning the handover to operations before the launch date is set. Both sides have to be specific enough to recognise on a Tuesday morning, in a meeting, under pressure.