Many responsible AI projects in organisations follow a familiar pattern. They start with momentum: powerful training, a newly signed policy, a motivated team — and then, a few months later, that momentum is gone. The policy sits on someone’s computer, but no one consults it; the training took place, but it did not become part of anyone’s routine; the initial enthusiasm faded as soon as it collided with the realities of day-to-day work. The challenge of cultural change around AI is not so much starting it as ensuring that it continues to live on once the novelty wears off. That is precisely when most projects begin to lose momentum.
Why cultural change fades, even when it starts well
Cultural change almost always fades for the same reasons, and none of them has much to do with the quality of the initial launch. It fades when it depends on a particular person who drives the change and then changes priorities or leaves the organisation. It fades when training is treated as a one-off event rather than integrated into everyday ways of working, the organisation’s own context and characteristics, and, above all, its values. It fades when a policy is drafted but never connected to real decisions, becoming a document that everyone signed but no one uses. And it fades when change is framed as a project with an end date rather than as a new way of working. A project ends. A culture, if it truly becomes part of the culture, does not.
The fatigue nobody talks about: choosing what and how, rather than trying to cover everything
There is a factor that is rarely mentioned when discussing why change fades: fatigue. Not physical fatigue, but the exhaustion of trying to keep up with the pace of technological change every week, every new tool you are expected to try, every update that changes everything, every headline announcing that yesterday’s technology is already obsolete. A team caught up in that race does not develop judgement; it develops exhaustion. And an exhausted team will disengage as soon as it can, no matter how well the project started.
The answer to that fatigue is not to run faster, but to choose better. An organisation does not need to adopt every tool that appears or master every new development. It needs to decide, based on clear criteria, what is genuinely useful and let go of the rest without guilt. Sustaining cultural change also means protecting teams from overload, giving them permission to ignore ninety per cent of the noise and focus on what genuinely adds value to their work. The ability to choose what matters and distinguish signal from noise is itself a form of judgement — and probably one of the most necessary when the novelty never stops coming.
What makes change last
What sustains change is not the intensity of the launch, but its integration into everyday routines.
A criterion for the use of AI remains alive when it becomes part of conversations that are already happening, when it appears in team meetings, in the onboarding of new people and in the way a tool is assessed before it is adopted. Instead of relying on an extraordinary momentum, it is embedded in the ordinary moments when decisions are already being made.
It also helps when the judgement does not reside in a single person, but is shared across the organisation, so that when someone leaves, the change does not leave with them. It also helps to have visible signs that what was agreed still matters: reviewing decisions against the agreed criteria, talking about what worked and what did not, updating the policy when reality changes, and assessing what impact all of this is actually having and whether it is genuinely worth the effort.
The role of review and ongoing support
Cultural change needs review points: moments deliberately built into the organisation to question whether what was decided still makes sense and is actually being applied. These do not need to be major evaluations. A simple periodic review can prevent the criteria from becoming a dead letter without anyone noticing.
And we should be honest about what happens when external support comes to an end, if there was any. Training or consultancy can ignite change and set it in motion, but it cannot sustain that change for the organisation. That is precisely what the organisation must be able to do once it has internalised the judgement to the point where it no longer needs to be reminded from the outside. The goal of any well-designed support programme is, precisely, to become unnecessary.
What is really being built
Behind the sustainability of cultural change around AI lies something much bigger than the adoption of a technology. What an organisation builds when it makes judgement endure is an ability to make sound decisions in the face of every new technology that emerges — and there will be many.
A company or institution that has learned to integrate AI with judgement and sustain that judgement over time develops, almost without realising it, the ability not to be carried away by every new development, to ask questions before adopting it and to decide rather than simply react. That ability is what truly protects an organisation in an environment that will not stop changing. AI is just the first test.
Frequently asked questions
Why do AI projects that started well fade?
Because they depended on one person driving the change, because the training was one-off and did not create lasting habits, because the framework or AI use policy was drafted but never connected to real decisions, or because the initiative was treated as a project with an end date rather than as a new way of working. And there is also the fatigue teams experience in the face of relentless technological change. A project ends. A culture does not.
How do I make sure our approach to AI stands the test of time?
Integrate it into existing routines — team meetings, onboarding, tool assessments — rather than relying on a one-off momentum. And share the criteria across several people, so that when someone leaves, the change does not leave with them.
How do I prevent my team from becoming exhausted by the pace of AI?
Protect them from overload. An organisation does not need to adopt every new tool or follow every development. It needs to decide, based on clear criteria, what is worth pursuing and let go of the rest without guilt. Giving the team permission to ignore the noise and focus on what genuinely adds value is itself a form of judgement, and it helps prevent the burnout that can ultimately undermine the entire initiative.
What happens when external support ends?
The organisation needs to sustain the change itself, and it can only do so if it has genuinely internalised the judgement required to make decisions. Good external support does not create dependency; it works towards making itself unnecessary. Its goal is for the organisation to know how to make decisions without needing someone from outside to remind it how.
