Working with AI Agents Exposed Something Major Transformations Often Miss

Over the past year, I have been working extensively with AI strategy and building AI agents to better understand where they create value, where they struggle, and what organisations need to consider as AI becomes increasingly embedded into day-to-day operations and decision-making.

What surprised me most was not the capability of the technology itself. It was how quickly AI agents exposed a challenge I have seen repeatedly throughout major transformation initiatives.

At the outset, objectives often appear clear, desired outcomes seem well understood, and stakeholders generally feel aligned around the direction being pursued. Yet when AI agents begin working with those objectives, subtle differences in interpretation often become visible surprisingly quickly.

As the work progresses, different views of success start emerging. Assumptions that had never been discussed become apparent, while context that seemed obvious to one group often turns out not to have been shared or understood by others.

AI agents simply exposed much earlier what might otherwise have remained hidden until later stages of delivery.

Alignment often exists around the objective, but not around its meaning, consequences and trade-offs.

Alignment often exists around the objective, but not around its meaning, consequences and trade-offs.

For organisations navigating major transformation complexities, this matters significantly. What appears to be alignment at the beginning can sometimes mask important differences in understanding that only become visible much later.

As momentum builds and activity increases, those differences can continue magnifying without attracting significant attention.

This is one reason why most transformation initiatives become harder than expected despite strong sponsorship, capable people and genuine commitment. The challenge is not always a lack of effort or intent, but a lack of shared understanding around the implications of what has been agreed.

AI is not creating that leadership challenge. It is simply making it much more consequential, much faster. The challenge of creating genuine shared understanding remains unchanged, but the consequences of getting it wrong are now being amplified at a scale and speed many organisations have never experienced before.

What assumptions are currently holding your transformation together that have never been consciously tested?