How Artificial Intelligence Exposed Corporate Think

Not just bad things come from the AI craze. It exposed what David’s slingshot is to the Goliaths of industry.
A while back I was working for an insurance company. During my tenure I had the pleasure to work in the AI and workflow automation department. Not a perfect fit on paper since I have a business degree. But my childhood obsession with video games sparked an even more intense love for computers, so it worked out great.
What was even more perfect was the timing: only a month after my arrival in the team, ChatGPT took the world by storm. I’ve never seen such a mania before. Our team went from the guys who automate boring work to the priests of the digital oracle of Delphi. We were bombarded with requests for LLM use cases. It was a Friday morning when this one request came. One from high up. Very high up. The C-Suite had honored us with their attention.
Someone in the upper echelons of the finance department had an idea. An idea that would revolutionize his department: “find me an AI use case for the finance department”.
We were ecstatic. The problem-solving approach might be backwards (here is the solution, find me a problem), but let’s be honest: that’s how innovation works sometimes.
Something felt off from the start. After digging for some potential applications for two weeks, we finally realized it. The real reason we were asked to find a use case. It wasn’t about improvement. It sure as heck wasn’t for the love of innovation.
It was about the story they get to tell.
And then the true tragedy unfolds.
Big corporations don’t exist. Not really. That would mean everybody in the organization follows the same goal. More accurately, big corporations are an amalgamation of smaller companies (departments) that all fight for finite resources (budget).
Which brings us back to the AI use case and the malady suffered by these companies. Sometimes, it’s more important to shout into the forest than to actually provide something of value. In our example, they wanted to implement an LLM-based solution where a simple script would do. This doesn’t just cost much more on compute, but is also more prone to errors because of its higher complexity.
It’s not the employee’s fault. The root cause lies in the incentive structure. Decision makers aren’t incentivized to make the best decisions possible, let alone innovate. If you only ride the wave of technology for signaling you’re innovative, your competitors who actually adopt the new technology will drown you.
My experience and these examples prove something to me again: even in our age of high technology, this remains a human world. It encompasses all our natural tendencies. And when money and status are involved, this gets turned up to eleven. A new technology might enable innovations beyond belief, but it won’t be used to its full potential if the incentives are not aligned. Or in the words of the legendary Charlie Munger:
“Show me the incentive, and I’ll show you the outcome.”
Even if the incentives don’t favor the best decision to make, make it anyway. Signaling value is a pop of smoke while creating value is concrete proof of skill. No wild approach needed. Problem first, solution second. That’s how you get noticed, not by anyone, but by the right people. That’s how corporate thinking turns into clear thinking.