Figma is one of the clearest examples of what happens when a successful software company recognizes the power of AI and decides it cannot afford to stand still.

As ChatGPT entered the mainstream and models such as GPT-3.5 Turbo and GPT-4o made increasingly visible leaps, the strategic pressure became obvious. This was going to be big. Companies such as Figma began connecting themselves to the magic box and integrating AI directly into their products.

Now imagine that you are an executive at a massive consumer-products conglomerate. The specific industry does not matter, but it helps to have something visual in mind, so picture a company like Johnson & Johnson, with hundreds of brands, divisions, suppliers, advertising relationships, and operational systems.

One day, you get a call from an old colleague at a competing company, Procter & Gamble. He tells you that his organization is aggressively implementing AI. Employees are worried about layoffs. Management is pressuring teams to use the new tools. AI adoption is becoming a central corporate priority.

You still do not know whether any of this will work.

But what if it does?

What if your competitor figures it out first and leaves you behind?

You decide that you should at least take a few meetings. You make some calls and sit down with your head of technology. That is when they present it to you.

It is the magic box.

You are flabbergasted. It feels as though you are staring directly into the future. The board is going to sing your praises. The magic box practically sells itself. You can deploy AI agents across the organization, automate repetitive work, analyze internal data, improve decision-making, and produce productivity gains in dozens of departments.

You have heard that AI is dangerous for software companies, but perhaps that is because software production is becoming a solved problem. Maybe traditional software stocks really are going to zero.

At least, that is what you begin to think.

A few months pass, and you are killing it. You need fewer employees. Results are improving. Costs are falling. Projects are moving faster.

Pop the champagne.

Then something else happens.

All those tiny operational problems that once required constant attention begin to disappear. The obscure exceptions, strange workflows, and niche problems that employees had patched together with band-aids are suddenly being handled by the system.

You sit back and watch your agent dashboard deliver reports from various areas of the business. The results are pretty good.

Then you speak to your old colleague from P&G again.

Strangely enough, his company has implemented many of the same changes as yours. He rants and raves about the efficiency gains and the stock price. You know it is not professional to get into specifics, but the opaque way he describes what is happening seems eerily similar.

Something else just seems off.

You speak to a web designer and ask for the Figma presentation on an app that is being launched. Then you notice Claude Design.

You realize that Claude ate Figma.

You scramble around researching the contracts you have signed. You realize that you are locked in and connected to this magic box, which learns about you and from the thousands of inquiries your employees have sent it.

You are now sharing intelligence with your competitors!

This seems awful. Is it?

And what's next?

Part II is forthcoming: it will look at what happens when the same handful of model providers sit underneath every competitor in an industry, and what a company can still control once its operational intelligence and its competitors' both run through the same box.