There are three types of companies in the AI industry right now. The ones building the models. The big platforms selling access to them. And a third group that almost nobody talks about, the ones quietly using AI to solve specific, unglamorous problems in specific industries and making a genuine difference in the process.
That third group is the most interesting one. And in my view, it's where the real value of this technology is actually being created.
Think about what that looks like in practice. A law firm that processes hundreds of contracts a week and has always relied on junior associates spending three days on due diligence. A hospital department drowning in patient data that nobody has the bandwidth to analyse properly. A bank's compliance team manually cross referencing transaction records against regulatory requirements that change every six months. These aren't exotic problems. They're the kind of problems that sit in the background of every serious organisation, quietly consuming time and money that could be spent elsewhere.
AI doesn't solve these problems by being impressive. It solves them by being consistent, fast, and available at a scale that no team of people can match. The technology isn't the point. Understanding the problem well enough to know how to apply it is.
This is where I think the public conversation about AI goes wrong. It focuses on benchmarks, on which model scored highest on which test, on who's winning the race to the biggest parameter count. None of that is particularly relevant to a logistics company that needs to reduce errors in its warehouse operations or a financial services firm trying to get better signal out of its reporting data.
What's relevant is whether someone understands both the technology and the business context well enough to build something that actually works inside a real organisation, with real data, real constraints, and real people who need to trust what the system is telling them.
That's a different skill set from building models. It's closer to what good BI and data work has always required: the ability to sit between complex data systems and the people who need to make decisions from them, and translate one world into the other without losing what matters in either direction.
The organisations that will get the most from AI over the next few years won't necessarily be the ones with the biggest budgets or the most advanced tools. They'll be the ones that found someone who asked the right questions before writing a single line of code.