“AI is useless.” It is the logical conclusion when you look at how most people use it.
Asking for a recipe with one leek, two eggs and some leftover cheese. Generating a poster for the football club barbecue, with a six-fingered goalkeeper. Rewriting an email so it sounds less passive-aggressive. Summarising an eight-minute video, because eight minutes is long.
And then concluding, very seriously: “Meh, it is overrated.”
Meanwhile, elsewhere
- A model predicted the structure of nearly every known protein, ending fifty years of deadlock in biology. It earned a Nobel Prize in Chemistry.
- Others design proteins that do not exist in nature, made to measure, to neutralise snake venom for instance.
- An AI found sorting algorithms faster than those written by humans. They are now in the C++ standard library: you run them without knowing.
- A neural network controlled the plasma of a fusion reactor, in real time.
- Researchers found new antibiotics by having a model screen millions of molecules.
- Scrolls carbonised by Vesuvius two thousand years ago, impossible to unroll without destroying them, are beginning to be read.
- Agents read a codebase, fix a bug, run the tests and open the change request, overnight.
None of this looks like a conversation.
The gap is the same in a company
We have a technology able to read papyrus burnt two thousand years ago, and we mostly use it to write promotional posters. The same gap exists in business. Giving everyone a chat window produces a few faster emails and a lot of disappointment.
The value appears elsewhere: an assistant connected to your own data, that answers from what your company actually knows. A tool that prepares a quote from a request and your price list. A step in a workflow that used to take someone an hour and now takes a check.
That is integration work, not a subscription. It needs your data, your rules and someone to decide what the model may and may not do.
What it means for you
The useful question is not “is AI useful?”. It is: in your business, which problem does everyone consider impossible, or simply inevitable, that nobody has thought of handing over to it yet? That is usually where we start.