Over the past couple of years, artificial intelligence has moved rapidly from peripheral experimentation into the centre of organisational attention. Leadership teams are being asked how prepared they are, boards are seeking reassurance, and programmes are being launched with the expectation that tangible productivity gains should follow.
The urgency is understandable. Few organisations want to appear slow or disconnected from developments that are shaping the wider market. As with earlier transformation waves, momentum builds quickly, and action begins even when understanding remains incomplete.
This pattern is not new.
Many leaders will recognise echoes of earlier Agile and product transformations. Those movements were rarely driven by superficial motives. They reflected a genuine desire to improve responsiveness, shorten feedback loops, and organise work more effectively around value. Yet in many cases, the outcomes fell short of expectations.
What became visible over time was not a failure of the models themselves, but the limits of the systems they were introduced into.
When transformation exposes discomfort
Agile transformations, in particular, surfaced uncomfortable truths. They made visible where decision-making was unclear, where accountability was blurred, and where leadership habits struggled to adapt to environments shaped by uncertainty rather than control.
For some organisations, this exposure proved productive. It created space for learning and adjustment. For others, it generated discomfort that was harder to sit with. The signals revealed by the transformation were softened, rationalised, or redirected into process discussions that felt safer than confronting deeper questions about behaviour, power, and capability.
AI now appears to be bringing those same tensions back into view, only at a much faster pace.
Where Agile gradually revealed limits in decision-making and learning, AI tends to surface them almost immediately. The speed of interaction, the visibility of outputs, and the ambiguity around responsibility leave far less room for avoidance. Patterns that once took months or years to emerge now appear in weeks.
The challenge is not that AI creates new organisational problems. It amplifies existing ones.
A familiar rush toward solutions
In many organisations, conversations about AI move quickly toward implementation. Questions arise about tooling, licences, data readiness, and rollout plans. These are important considerations, but they often arrive before a more fundamental conversation has taken place.
What is the organisation actually trying to use AI for?
In practice, this question is frequently unanswered. The desire to keep pace with market trends can overshadow the need for clarity of intent. Leaders know they must act, but struggle to articulate where AI should meaningfully support work, decisions, or outcomes.
As a result, preparation begins before purpose is defined.
This dynamic is strikingly similar to earlier transformations. Teams were trained in Agile practices before understanding what problems those practices were meant to solve. Product roles were introduced without clarity about decision rights. Structure arrived before sensemaking.
AI risks repeating this sequence, with greater complexity and higher stakes.
Capability and the reality of use
Once specific use cases are identified, a different set of questions emerges. Not whether AI can be deployed, but whether it can be used well.
Effective use of AI depends on how people think, decide, and exercise judgement in real situations. Capability emerges at the intersection of judgement, context, and environment, and becomes visible most clearly when conditions are uncertain and consequences matter.
This is where many organisations encounter friction.
AI often assumes that people can frame meaningful questions, interpret outputs critically, recognise limitations, and remain accountable for decisions that are partly informed by machines. These expectations may never have been made explicit, let alone developed deliberately.
When those capabilities are uneven or unsupported, behaviours begin to shift in predictable ways. AI is used cautiously and tactically rather than confidently and purposefully. Experimentation moves into the shadows. Value remains difficult to evidence. Concern about risk grows, sometimes out of proportion to actual exposure.
These patterns do not signal resistance. They reflect a system navigating uncertainty without clear guidance.
Patterns that tend to emerge
When capability is not addressed deliberately, several familiar dynamics appear.
AI usage becomes fragmented across teams, creating inconsistency rather than shared learning. Conversations drift toward adoption metrics instead of outcomes. Leadership confidence wavers as early enthusiasm gives way to ambiguity. Over time, the initiative begins to feel heavier rather than enabling.
What is often interpreted as an AI problem is, in reality, a capability and system issue.
The same was true in earlier transformation efforts. Models introduced into environments unprepared to support them did not fail outright. They stalled, stretched, and gradually lost meaning.
AI is exposing those same limits again, with less tolerance for delay and far greater visibility.
A more grounded place to begin
A more constructive starting point sits one step earlier than most current conversations.
Before discussing tools or readiness, organisations benefit from identifying the specific use cases they want to pursue. What kinds of work, decisions, or outcomes are they seeking to improve? Where does AI genuinely have the potential to assist rather than distract?
Only then does the more meaningful question arise. What capabilities must exist for those use cases to function responsibly and productively within this organisation?
This framing anchors AI in purpose rather than trend. It links technology to real work. It also makes visible where development, support, and leadership attention are required before scale becomes viable.
In doing so, it shifts the conversation from implementation to readiness in its fullest sense.
Closing reflections
What is becoming clearer is that the challenge organisations are facing is less about introducing new technology, and more about whether their existing ways of thinking, deciding, and learning are able to support it. As with earlier transformations, outcomes will be shaped far more by the human system than by the sophistication of the tools themselves.
As you reflect on your own context, it may be worth pausing to consider a few questions.
- Where is your organisation already experimenting with AI, and how clear is the intent behind those experiments?
- Which decisions are being supported by AI today, and how confident are you about accountability when outcomes are uncertain?
- What capabilities does your system currently reward, and which ones effective AI use actually requires?
- Where might urgency be crowding out understanding?
These questions do not demand immediate answers. Their value lies in what they help reveal.
I will return to these themes in future reflections.
More to come. Leave your thoughts and comments below.
Leave a Reply