AI is usually described as a technological revolution. I think it is fundamentally a human journey. Technology is advancing quickly, but the value organisations realise depends on how fast people develop the judgement, discipline and accountability to use it well.
This journey applies primarily to what I call Horizontal AI or productivity scenarios with general-purpose tools such as Copilot, ChatGPT and Claude. They offer enormous flexibility, but broad outputs require strong human judgement, domain knowledge and verification. The benefits usually go to individuals and small teams. Usually, when team members leave, the value can be lost. This adds tactical value to organisations. The freedom of horizontal AI creates creative upside without requiring compliance knowledge or domain expertise. That same freedom is why behavioural guardrails and accountability matter so much.
The other area is what I call Vertical (system thinking) AI: It is built around a specific domain, process or problem. It’s usually a system-thinking AI use case, which you can find more about here: https://lnkd.in/p/ennQX8Zr.
From my own experience and from talking to people adopting tools like Copilot, ChatGPT and Claude, I see five stages. Some people move through them quickly. Others get stuck. The challenge for organisations is not encouraging adoption. It is helping people move through the stages safely enough to turn activity into real productivity.
Stage 1: The “Whoa” Stage
The journey starts with amazement. You ask a question, upload a document or request an email and within seconds something useful appears. The experience feels powerful because the interface is simple, the response is immediate and the possibilities seem endless.
People here are curious but still cautious. They experiment, test the boundaries and usually check whether the outputs are accurate. That combination of curiosity and scepticism makes this a healthy and necessary phase.
Typical behaviors:
- Exploring different tools and use cases
- Asking lots of questions
- Testing outputs against what they already know
- Discovering creative possibilities
Stage 2: The Addiction Stage
Curiosity becomes habit. AI starts appearing in emails, reports, research, presentations and everyday problem-solving. Prompting feels productive because content and ideas arrive almost instantly.
This stage produces genuine benefits, but it also blurs the line between activity and productivity. AI makes it easy to generate more information than anyone can absorb, which creates new work in reviewing, refining and validating the output.
Typical behaviors:
- Using AI for almost every task
- Generating more content than is needed
- Spending too long refining prompts
- Struggling to start work without AI
- Mistaking more output for better outcomes
The addiction is not to AI itself. It’s the feeling of acceleration.
Stage 3: The Hallucination Stage (Risky Stage)
The greatest risk appears when confidence grows faster than judgement. We talk constantly about AI hallucinations, but humans hallucinate too, trusting polished and confident outputs without properly challenging them.
AI generates information faster than people can process it. That creates cognitive overload and makes it harder to separate genuine insight from convincing noise. People feel highly productive while the quality of their decisions quietly declines and the workload of reviewing or fixing shifts to others. Overall organisational productivity might lower.
At this stage people often:
- Trust information without verifying it
- Confuse confident language with accuracy
- Accept recommendations without challenge
- Become overconfident in unfamiliar domains
- Lose sight of the original problem
- Decide using incomplete or misleading outputs
Domain experience helps. Someone with grounded knowledge is more likely to spot a missing assumption or an unrealistic recommendation. But experience is not immunity. Experienced people also become over-reliant, particularly when the output confirms what they already believe.
This is the most dangerous stage because people rarely realise they are in it.
Stage 4: The Insight Stage
The most important organisational transition is moving people from hallucination to insight. This is where risk reduces and sustainable productivity begins.
Here, people stop using AI habitually and start using it intentionally. They understand where it adds value, where it adds noise and where doing the task directly is simply faster.
Mature users:
- Ask clearer, more focused questions
- Request shorter, more relevant outputs
- Challenge recommendations
- Verify anything consequential
- Apply their own domain knowledge
- Use AI selectively rather than automatically
People at this stage often appear to use AI less. They have not rejected it. They have developed the judgement to know when a conversation, personal experience or a self-written email produces a better result. The technology has not changed. The human has.
Stage 5: The Trust Stage
Once people understand where AI fits, organisations can embed it responsibly into workflows, processes and operating models. The question shifts from “How can I use AI?” to “How should people and AI work together to achieve a better outcome?”
The trust stage is not blind confidence. It is understanding where AI is reliable, where it can fail and where human judgement must take priority. At this stage, some scenarios or use cases can move into Vertical Systems Thinking AI use cases.
At this stage, the person focuses on:
- Redesigning work rather than accelerating tasks
- Automating repeatable, appropriate activities
- Focusing on review and governance
- Managing information contextualization, making sure they don’t burn out or burn people out with too much information.
- Keeping human review around important decisions
- Making outputs traceable and challengeable
- Maintaining clear human accountability
AI becomes part of the normal operating environment rather than the center of every conversation. The strongest sign of maturity may not be using AI more, but knowing exactly when to use it, when to question it and when to switch it off.
Moving People Through the Journey
There is no magic bullet. People bring different levels of experience, confidence, curiosity and scepticism. Everyone experiences the “Whoa” Stage. Many move into addiction. Some stay in the Hallucination Stage longer than they realise, which can lead to more risk than benefits. Security, privacy and compliance controls matter, but organisations also need decision-making guardrails that define:
- When AI should or should not be used
- Which outputs need verification
- What evidence is needed before acting
- When domain expertise should override AI
- When to involve experts instead of acting as one
- Who owns the final decision and outcome
Accountability cannot be transferred to a model. Anyone using AI for advice, analysis or recommendations remains responsible for the result.
The organisations that create lasting value will help people reach insight and build trust quickly and responsibly. AI maturity is not about how often we use the technology, but how wisely we choose to use it.
