AI makes it easier to move fast in the wrong direction

In health and pharma, judgement is both critical and difficult to balance. Commercial, medical, customer, compliance and technology all bring a valid perspective on what should shape the experience. The challenge is that these perspectives do not always point in the same direction, and alignment can easily be mistaken for clarity.

We are all familiar with the failure rates attached to technology programmes, and AI project failure rates appear higher still. There is always a cost to failure - financial cost, lost time, missed opportunity, and often customer impact too. That impact is not always easy to measure, but it is felt in every interruption that adds little value, every poor experience that creates friction, and every moment where trust is weakened rather than earned.

When the results are reviewed later, the focus is usually on the solution itself. And that is understandable, because the solution is what people can see. But in most cases, the first mistake did not originate there.

Most teams are not short of ideas. They are short of precision about what deserves to be solved. In many disciplines, this is referred to as the problem statement. In product teams, it becomes a 'user problem' or 'job to be done'. In consulting, it is the key issue to be resolved. Different language, same intent.

To define, with enough clarity, what is creating friction, for whom, at which moment, and with what consequence - and why solving it creates a meaningful commercial advantage. Not in abstract terms, but in a way that can guide real decisions.

Without that level of definition, teams might commit to a direction that feels credible, align around it and execute it with confidence. But if the problem is off, everything downstream can perform exactly as designed and still deliver very little.

Without proper exploration and real clarification of the problem statement, organisations tend to do one of two things. They either move forward using existing narratives, or they delay in an attempt to remove uncertainty entirely. Neither resolves the underlying issue. And when the issue remains poorly defined, the opportunity is rarely realised in full.

AI makes this distinction more important, not less. It can help teams synthesise insight and move faster, but it does not correct weak foundations.

If a team is aligned around the wrong problem, AI will help them execute it more efficiently and at greater scale. If a team misunderstands behaviour, AI will personalise that misunderstanding. It does not create clarity. It amplifies whatever level of judgement you start with.

As the cost of producing solutions falls, the cost of defining the wrong problem rises. You see it in services that look coherent internally but fail to change behaviour, in experiences that scale technically but are ignored, and in investment that moves quickly but delivers little measurable impact.

In a fixed pre-launch window, there is limited opportunity to correct course. You do not get multiple attempts at defining the problem. You get one.

The distinction is not really about methods. It is about risk.

In some cases, the primary risk is alignment risk. The organisation may already hold enough insight to define the right challenge, but that insight is fragmented, diluted, or interpreted differently by different functions. The danger here is not the absence of knowledge. It is that teams move forward without a shared view of what matters most. In that environment, speed can be useful, because the real need is to interrogate what already exists, expose contradiction, and create focus.

In other cases, the primary risk is understanding risk. The organisation may be aligned internally, but aligned around an incomplete or inaccurate picture of customer behaviour. The danger here is not confusion. It is false confidence. Teams believe they understand what is driving hesitation, drop-off or inaction, when in reality the most important signals are either missing or misread. In that environment, moving quickly only compounds the problem, because execution begins before the behaviour has been properly understood.

The important point is that these are different risks, and they require different responses. One asks whether the organisation is aligned. The other asks whether the organisation is right.

Both routes lead to the same outcome - a clearly defined, high-value challenge and credible solutions. But they solve different problems. One addresses whether the organisation is aligned. The other addresses whether it is right.

This is where judgement matters. Instead of asking whether to move fast or slow down, a better question is: what is most likely to be wrong in our current understanding?

If the issue is internal confusion, speed helps. If the issue is behavioural blind spots, speed hurts. A more effective approach is to start with what you know, test how far it holds, and introduce real-world validation where confidence breaks.

The advantage does not come from speed alone, or from depth alone. It comes from knowing when existing insight is enough, and when it is not.

To move fast, some assumptions will always be required. But many organisations carry more assumptions than they realise, and move with more confidence than the evidence allows.

Closing that gap is not insurmountable, nor does it need to take long. This is increasingly where we are helping clients. Not starting with solutions, but helping teams define the problem statement with enough precision to guide real decisions. In some cases, that means pressure-testing what already exists. In others, it means bringing real customer behaviour into the room early, so assumptions are challenged before they are scaled.

When that foundation is right, something shifts. Decisions become easier. Trade-offs become clearer. Teams move with a shared sense of direction, rather than negotiating competing interpretations of the problem.

AI then becomes genuinely valuable. Not as a starting point, but as a multiplier. The ability to diverge and create more options, move faster, test more, and then execute with confidence.

Speed then comes with confidence, because the direction is clear and teams are aligned.