The paradox is that as AI platforms get more advanced, what stands between them and the enterprise turns out to be the human factor.

In practice every startup building AI today faces the same problem: how do you build a product that genuinely fits every customer, when every organization expects something tailored to its own shape.

Over the past few weeks we hosted at Grove Ventures the executives leading Forward Deployed Engineering in these companies, from public companies to twenty-person startups. The motion is showing up in more and more places, under different names. And without exception, one thing repeats: looking back, all of them wish they had started it earlier.

Palantir invented the role twenty years ago. For most of that time almost no one else adopted it. Now it is becoming something every company of this kind needs. In the US, the number of these roles jumped ninefold in a year.

We call it the FDE paradox. The more AI advances, the more people its deployment requires, not fewer.

We will write about this in more depth soon. In the meantime, if you are building an enterprise AI company and still treat FDE as a support function rather than a growth engine, now is the time to change that.