Many Data and AI Programs Plan for Delivery. Few Plan for Behaviour Change

Most data and AI programs are built around use cases. Deliver the capability, prove the value, move to the next one. Strategy gets documented, governance gets designed, and delivery roadmaps get built and tracked.

What almost nobody plans for is whether people will actually change how they work.

That omission is not trivial. It is usually the reason programs that look successful on paper produce disappointing results in practice.

Rational people, wrong conditions

This is not always about people being resistant. It is people being rational.

A commercial team has access to a pricing model that optimises margin, but they continue to set prices based on their own judgement. Not because the model is poor, but because they are measured on short-term revenue targets and experience has taught them how to hit those targets reliably. The model introduces uncertainty, and in that context, ignoring it is the sensible choice.

A supply chain function invests in demand forecasting analytics, and planners continue to rely on historical patterns and manual adjustments. Not necessarily because they distrust the technology, but because if the model is wrong they still own the outcome, and following their own judgement at least gives them an explanation.

In both cases, the organisation introduced new capability without changing the conditions under which decisions get made. The behaviour that followed was entirely predictable.

What gets missed at the top

When behaviour design is absent from a program, it shows up at leadership level too.

Senior leaders articulate genuine ambition around data and AI, but when the pressure is on, decisions still get made on experience and instinct, analysis gets requested selectively, and data gets used to support positions that have already been taken.

Nobody designed the reinforcement loops. So the old behaviours persist at every level, including the ones the rest of the organisation is watching most closely.

The discipline that is almost always missing

Behaviour is shaped by what people are measured on, what they are rewarded for, and how easy it is to do the right thing in the moment a decision gets made. These factors operate continuously in the background and they do not wait for the program to catch up with them.

BJ Fogg from the Stanford Behaviour Design Lab, developed a model that makes the mechanism explicit. Behaviour occurs when Motivation, Ability, and a Prompt converge at the same moment, and if any one of them is missing the behaviour doesn't happen.

Applied to data and AI adoption, the diagnostic is straightforward. Are people actually motivated to change how they work, based on how they are measured and rewarded? Is it easy enough to do within the flow of real work, or does it require extra effort on top of an already full day? Is there a clear trigger at the moment the decision gets made, or does the new way of working exist somewhere off to the side, available but never quite activated?

In most organisations at least one of these is weak, in many all three are, and in almost none has anyone deliberately designed for all three together.

How Q22 approaches this

Q22's Levers and Incentives Map, one dimension of the SCALE© diagnostic, makes the behavioural layer visible and actionable across three areas.

Decision habits looks at how decisions actually get made under time pressure and uncertainty, not how they are described in process documents, and where data gets used versus where experience and instinct still dominate.

Friction and ease examines how accessible data and AI tools are within the flow of real work. Motivated, capable people will still avoid tools that are slow, disconnected, or require a context switch to use, and friction is often the simplest thing to fix and the last thing anyone looks at.

Reinforcement loops identify what behaviours the organisation is actively reinforcing through its formal mechanisms and through the informal signals leaders send when the pressure is on, two things that rarely say the same thing.

Where to start

Not with more communication, and not with another training program.

With an honest look at what the organisation is actually incentivising today: what gets rewarded, what gets avoided, and what signals leaders send through their own behaviour when it counts.

That picture is usually clear once someone looks for it.

When friction is reduced, incentives are aligned, and the right behaviours are visibly reinforced by the people with the most influence, things begin to shift.