Most companies aren’t becoming AI businesses. They’re becoming businesses that use AI.

On a recent podcast a colleague of mine was recounting the time he was invited to participate in an FCA and Bank of England forum looking at AI in financial services. One of the first things he discovered was that nobody could agree on a definition of AI.  

To some it meant automation; others saw it as a rebrand of machine learning; and others thought it was to do with neural networks. A room full of smart people and all with different perspectives on AI. But that’s not a reflection on the participants, and instead illustrates the problem with a simplistic ‘AI’ label. AI is a category after all, not a thing in and of itself.  

So everyone was using the same term, but they weren’t necessarily having the same conversation. Hold that thought.

Over the past couple of years we’ve been told that every business needs to become an ‘AI business’. Conferences have been built around it. Investors ask about it. Boardrooms discuss it. And some consultants have done very well out of it. 

Yet when you look at what’s actually happening inside most companies, very few are building foundation models or creating entirely new AI products. Most are trying to work out how AI fits into the work they already do. 

A surprising number of companies seem to buy the AI tool first and only then start looking for somewhere to use it. That’s understandable as the technology is improving so quickly that it’s easy to feel you’re being left behind if you’re not experimenting. Updates come out so quickly that once you’ve mastered one another has usurped it, including models too dangerous for the public being released to the public.  

But starting with the technology often leads to disappointment. The more useful starting point is much less exciting, and it’s about establishing where AI might actually be helpful. For example, where are people currently wasting time? Where are customers getting frustrated? Where does work get stuck? Those questions aren’t particularly sexy, but they tend to reveal where the real opportunities are. 

People don’t buy AI  

Our customers don’t come to us because they’re excited about AI. They come to us because they want a faster, simpler way to buy a car. 

That sounds obvious, but it’s surprisingly easy for businesses to lose sight of. Because when you’re inside an organisation, it’s natural to become fascinated by the technology. Builders are by their nature attracted to the shiny new object. Customers tend to stay focused on the outcome. 

Nobody buying a car cares what machine learning model sits behind a lending decision, or what painstakingly developed and trained AI assistant has helped them get unstuck. They care whether they can get a fair answer quickly, whether the process is straightforward, and whether they trust the company they’re dealing with. 

The technology only matters if it helps deliver those things. 

AI is less technology and more culture 

Which takes us back to that forum. The disagreement was never really about the technology at all. That’s one of the reasons I’ve become sceptical of the idea that AI adoption is primarily a technology challenge. In many cases, it’s a challenge about how work gets done and, importantly, how a business believes work should get done. 

Take analytics as an example. In lots of organisations, data teams gradually become bottlenecks without anyone intending them to. Every report, every dashboard and every question flows through a relatively small group of specialists. As the company grows, demand grows with it. 

The result is that highly skilled people spend a surprising amount of time answering routine questions, but AI can help change that.  

Not because it removes the need for expertise, but because it allows more people to answer simple questions for themselves. The specialists are still there. They’re still essential. They just spend less time acting as a reporting service and more time tackling problems that genuinely require their experience.  

So again the question becomes, how does a business believe work should get done? Everything through a human, everything through AI, or a more balanced ‘human in the loop’ approach that places human expertise and judgment at the core, supported by technology that enables people to focus on the areas where they can provide most value? 

That may not sound like a revolutionary use of AI but the reality is that many of the most valuable applications aren’t revolutionary at all, they’re merely practical. They remove friction, save time, eliminate repetitive work that no-one likes doing, and those gains compound massively. 

Use case: AI in lending 

The same principle applies in lending. One of the most powerful uses of AI isn’t necessarily making decisions faster. It’s making better decisions. 

Credit has always been a balancing act. Lenders need to understand risk, but they also need to make sure they’re not unnecessarily excluding people who can afford to borrow responsibly. Better data, better models and better analysis can improve those decisions. And that’s where AI becomes genuinely useful. 

Not because it’s AI, but because it helps solve a real problem. It’s how we’ve developed our latest decisioning model – Senna. Built using more than one million data points from six years of industry data, the model uses gradient boosting machines to improve predictive accuracy. And since we launched the model we’ve seen approval rates rise by 5%, and the best-performing half of applicants now represent 25% lower risk compared with the previous model. 

AI is a tool. It doesn’t remove bad workmen 

I don’t think the companies that benefit most from AI will necessarily be the ones with the biggest budgets or the most ambitious announcements. 

Technology advantages rarely stay exclusive for very long, and eventually everybody has access to broadly similar tools because if it can be bought, it will be bought.  

So what matters more is whether an organisation that’s investing in AI has the culture and talent that can truly make use of the tools.  

The businesses that see the greatest returns from AI are the ones full of curious people who are willing to question why a process exists, whether a task still needs to be done manually or how a problem could be solved in a completely different way. The sort of people who see something that needs to be done, and find a way to do it when there’s no existing playbook. 

A lot of businesses are still asking how AI can transform their company, but the better approach is to find out where time is being wasted by skilled people and then give them the tools they need to automate, expedite and enhance their work.  

If you give a member of the public the best football boots ever made, they still won’t perform as well as Erling Haaland with both legs tied together. So the real power is talent density supported by the most appropriate AI technology. 

AI with the right culture is where success is found 

The businesses that succeed won’t necessarily be the ones talking about AI the most. They’ll be the ones quietly using it to remove friction, make better decisions and give their people more time to focus on work that actually matters. 

Most customers will never notice which parts of those businesses are powered by AI and which aren’t. And that’s probably a sign they’re doing it right.