AI is moving faster than anyone can track. So how do you keep up?
Most business leaders are not prepared for what is coming. Not because they are not intelligent or not working hard, but because the cascading implications of AI adoption are genuinely difficult to see clearly when you are inside a business trying to run it day to day. The technology is moving fast. The economic signals are mixed and often contradictory. And the people whose job it is to advise business leaders - accountants, lawyers, banks, consultants - are largely working from the same incomplete picture.
That is the problem I have been trying to solve.
Someone asked me yesterday how I keep up with AI and maintain a clear view on where the world is going. I did a terrible job of answering it in the moment. I have thought about it a lot since, and the answer is more interesting than the one I managed to give him.
Look at what the data is telling us right now. Youth unemployment has hit 16.2% this week - the highest in over a decade. Real wage growth has collapsed to 0.1%, with the ONS flagging it is likely to go negative as the energy price shock feeds through. Vacancies are at a five-year low. Consumer confidence nudged up two points in May to -23 but remains four points lower than this time last year. That is not a recovery. It is a fragile bounce in a deteriorating picture.
The question is whether you read this as cyclical noise or something more structural. I think it is the latter. And the reason I think that matters enormously for every business leader trying to make decisions about their workforce, their cost base, and their competitive position over the next three to five years.
Here is what I think it actually takes to build a view worth acting on.
You need institutional analysis from the top. The IMF and the World Economic Forum have both done serious work on the scale of potential labour market displacement - the WEF Future of Jobs report and the IMF's World Economic Outlook are the two I return to most. They are not infallible and some of their assumptions deserve challenging, but they give you the macro frame and the historical context. Pulling that research together for an AI investment thesis I have been working on - stress-testing the numbers, working out where the institutional consensus holds and where it falls apart - was itself an education.
But institutional research only takes you so far. It tells you the scale. It does not tell you what is actually happening inside UK businesses, on the ground, right now.
You need operational experience. Not reading about businesses going through change but having been inside one at the level where the decisions actually get made. Twelve years running a manufacturing business gave me a calibrated sense of the gap between how organisations say they will respond to disruption and what they actually do. That gap is large and it is consistent. Anton de Leeuw reinforces this from a different angle - he restructures mid-market businesses internationally and what he sees inside companies under real financial pressure across multiple geographies confirms what I have seen in smaller businesses closer to home. The human resistance to change, the lag between a leader's stated intention and their organisation's actual behaviour, is one of the most important and least discussed variables in any serious analysis of AI adoption.
You also need a read on where the technology actually is, as opposed to where the headlines say it is. Through SETsquared, Angel Investors Bristol and working groups on deploying capital in the AI era, I spend time regularly with people building early-stage AI businesses. They work on specific problems - gaps in current capability, the distance between what a tool can do in a demo and what it can do reliably inside a business running under normal conditions. Jim Moodie sharpened my thinking here considerably. As a tech leader and investor, his read on how technological capability actually translates into economic reality is as precise as anyone I have encountered on this journey.
You need the mid-market signal. Joanne Rumley at Foot Anstey, for example, operates at the intersection of large business strategy and legal reality - a corporate lawyer who advises at board level, her read on how organisations actually respond to structural change is grounded in the conversations happening right now in the most consequential rooms. When she talks, I listen.
And you need to use the tools yourself. Not as a novelty but as a genuine stress test of what they can and cannot do under real conditions. There is no shortcut here.
What none of these sources does on its own is give you the full picture. The institutional research tells you the scale but not the texture. Operational experience tells you the texture but can lead you to over-index on what you have personally seen. The founder ecosystem is sharp but optimistic by nature. The professional network is broad but cautious. Each corrects for the others. The people I have mentioned were not in the same room - I encountered each of them separately across different parts of this journey - but the collision of their perspectives is where I form my view. I 'keep up' through relentless reading, having conversation after conversation with extremely capable people, and through immersing myself in dynamic and undynamic businesses.
One gap I will own honestly. I am not a regulatory analyst and I have no particular feel for where UK and EU policy is heading in practice. The policy layer is missing from my picture. If you work in that space and are actually doing something to prepare for the world we are slowly moving towards - rather than standing on the bridge like Captain Smith watching the iceberg approach (yes, I have been listening to The Rest Is History!) - I would genuinely like to speak to you.
What I have built, alongside the working group, is a model tracking three scenarios for how AI adoption plays out across the UK labour market: a co-pilot economy where human and machine augment each other productively, a long rotation where displacement runs ahead of retraining and the labour market adjusts slowly and painfully, and a permanent split where a structural divide opens between those who can work alongside the technology and those who cannot. The model is updated quarterly against ONS and GfK data. The numbers I opened with - youth unemployment at a decade high, real wages flatlining, vacancies contracting, consumer confidence fragile - are the observable signals the model told us to watch. They are not pointing at the co-pilot economy.
[LINK - interactive model: track the scenarios as the data comes in]
The direction of travel is clear enough that I am now building the next stage. The investment thesis mapped the macro picture. What business leaders actually need is something more practical - a structured way to assess their own exposure and act on it. I am developing a working session that takes the macro analysis, the defensibility work and the economic scenarios and turns them into something tangible: an adoption imperative, a cascading risk assessment, a defensibility audit, a foundational diagnostic and a transformation roadmap. The goal is a session that resonates with the people running established businesses, not just those evaluating them from the outside. I will be sharing this work at private events in the coming months. If you want to be in the room, get in touch.
The question yesterday was a good one. Not because I did not have an answer, but because I had never been asked to lay out the method rather than just the conclusion.
If you are reading this from a different country or a different sector and you are seeing something I am not, I would like to hear it.