← All writing

AI & Society

Why Most Companies Are Approaching AI Backward

They start with the technology and go looking for a problem. It almost always works better the other way around.

Here is a conversation happening in a lot of companies right now.

A leader says they need an AI strategy. Someone asks what problem they are trying to solve. There is a pause. Then you hear about a competitor, or a board member, or an article, and somewhere in there is the real answer, which is that they are afraid of being left behind.

That fear is reasonable. It is also a terrible place to start.

Technology first, problem later

The backward approach goes like this. Pick a model. Pick a vendor. Find a team to "do AI." Ask them to find places to use it. Launch a pilot. Announce it.

Six months later the pilot is still a pilot, nobody uses it much, and the people who built it are quietly frustrated because they were never told what success looked like.

This is not an AI problem. The same thing happened with mobile, with cloud, with big data, with blockchain. I built through several of those waves, mobile advertising at Receptiv, connected TV at BrightLine, cloud transformation at LivePerson. The technology changes. The mistake does not.

Problem first, technology last

The version that works starts somewhere unglamorous.

What is slow? What is expensive? Where do customers get stuck? Where are your people spending hours on something they hate and that nobody would miss if it went away? Where are decisions being made on bad information because good information takes too long to gather?

Write those down. Put a number on them if you can. Then, and only then, ask whether AI is the right tool for any of them.

Sometimes it is. Sometimes the answer is a better spreadsheet, or a process change, or one honest conversation between two departments. That is fine. The goal was never to use AI. The goal was to fix the thing.

Three questions before you build anything

What changes for the person on the other end? Not the efficiency gain on a slide. The actual human, a customer or an employee, and what their day looks like after.

How will we know it worked? If you cannot measure it, you cannot tell a good pilot from a bad one, and every pilot will look promising forever.

What happens when it is wrong? AI systems are wrong in ways that can sound very confident. Someone needs to own the failure case before you ship, not after.

Start smaller than feels impressive

The best technology projects are almost boring when they start. One workflow. One team. One clear measure. They work, people trust them, and then they grow.

The worst ones start with a press release.

The fear is still real

I do not want to wave away the fear of being left behind. It is a reasonable read of the market. But the companies that will actually be left behind are not the ones who started late. They are the ones who spent two years building things nobody needed, and lost the trust of their own people doing it.

Start with the problem. The technology will still be there when you get to it.

If your organization is in the middle of this, this is exactly the work I do as an advisor. I am happy to talk about where you are.

AI strategyproductleadershipadvisory
Jaime Boyaca

Technology executive, speaker and explorer of consciousness. This idea is also the heart of his keynote From Reaction to Response.