Approach
Getting more from AI starts with seeing the capability more clearly. Prompts and tools matter, but they become much easier to use well once you have a useful mental model of what sits behind them.
Understanding helps you work with AI more deliberately: recognising where the interaction is falling short, changing direction and seeing possibilities you might otherwise miss.
The aim is to connect that capability to the knowledge, standards and ambitions you already bring. That movement from understanding to fluency to application is at the centre of the approach.
Start with understanding
AI is unusually easy to use before you understand it. You can open a chatbot, ask a question and get something impressive back within seconds. That accessibility is part of its power, but it can also leave people badly wrong-footed about what is happening underneath.
Without a useful picture of what kind of system you are dealing with, it is easy to treat every good answer as evidence that it understands more than it does, or every strange answer as an inexplicable failure. Interaction can become trial and error, and increasingly sophisticated use can still amount to a collection of tricks rather than a coherent way of working.
Approachable AI uses clear explanation, analogy and demonstration to make the underlying technology graspable: how these systems behave, where their strengths lie, where they are unreliable and why different ways of working with them produce different results.
With that foundation, the technology becomes less mysterious and more workable. You can recognise patterns in its behaviour, understand why context and direction matter, and make better sense of weak or surprising output.
Build fluency
At first, working with AI can feel like trying to discover the right combination of words. You change the prompt, try again and hope the next answer is better.
Fluency changes that. You become less dependent on finding the “right” instruction and more able to shape the interaction as it develops: adding context, introducing ideas, asking for challenge, comparing approaches, changing direction or deciding that the system has taken the work as far as it usefully can.
An important part of that is learning to diagnose what isn't working. Rather than simply rejecting a weak response and starting again, you become better at recognising how it misses the mark: the emphasis is wrong, the thinking is too generic, the structure works but the tone doesn't. You can then explain what needs to change and judge whether the next response addresses it.
As that capability grows, you can also vary the relationship according to the task: sometimes asking for straightforward assistance, sometimes developing an idea through dialogue, sometimes delegating a bounded piece of work, and sometimes using it to challenge your assumptions or red-team an emerging idea.
There is no single correct way to work with AI.
The capability lies in being able to choose, adapt and direct the interaction intelligently.
See more of what might be possible
Once you begin to see AI as a general-purpose technology, a wider range of possibilities comes into view. It can help you explore, compare, simulate, challenge, combine, analyse, create and communicate.
Seeing AI used in an unfamiliar way can change your sense of what you could attempt. In my training and consultancy, I develop relevant examples and live demonstrations to make that breadth tangible.
The value lies in recognising which of those possibilities make sense in your own work, expertise, context and ambitions — including things that may previously have been difficult, impractical or simply not visible.
Connect possibility to real work
What are you actually trying to achieve? What knowledge and judgement does the work require? Who or what is it for? What standards or constraints apply? Where might AI add something — and where might it not?
Those questions help establish whether an idea is worth pursuing. Just as important is what introducing AI would mean for the people involved and how things are currently done.
Approachable AI uses structured frameworks, including the Generative AI Canvas, to help people decide what to try, how to test the approach and what would count as a useful result.
Keep your expertise active
Working well with AI makes your own knowledge and experience more important.
The better you can frame the problem, explain the context, recognise quality, notice what is missing and judge what matters, the more intelligently you can direct what AI contributes. These are familiar professional capabilities, now being applied in a different kind of interaction.
That also means keeping critical thinking active: using AI to explore ideas, test assumptions and widen your perspective without becoming over-reliant on whatever the system produces.
You bring the purpose, expertise, standards and judgement that give AI’s capabilities direction. Responsibility and accountability for how the resulting work is used remain with you.
Using AI well also means knowing what information you can safely use in different systems, which organisational policies or requirements apply, and what those mean for the task in front of you. The aim is to help people make the most of AI’s capabilities while making informed, responsible choices about how they use them.
I help make AI clear enough to work with, identify where it could add value and develop the strongest ideas into practical applications.
The aim is not for me to remain ‘the expert’ at the centre. I want the people I work with to develop the understanding and capability to make their own judgements and direct the work themselves.
Work together
Approachable AI can meet people at different starting points, from building shared understanding to putting AI to work on a particular task or taking existing capability further.
Explore ways to work together ↗