Can AI Agents Have Human Instinct?

If you didn’t go to an elite college and work at a FAANG company, an AI agent has probably already screened you out of the ‘promising founder’ list.

I’ve had a string of catch-ups with other VCs lately. The hottest topic? Using AI agents to scrape LinkedIn and power programmatic outbound to ‘promising’ founders.

Criteria is all similar: elite school → big tech experience → modest prior exit → hot sector.

While I understand the pressure to ‘cover more ground’ and the temptation to let AI take the first pass, this approach is totally backwards for seed investing. When everyone optimizes for the same signals, non-obvious founders and out-of-favor categories get screened out before the first meeting. When the human element gets lost, the breakout companies get missed.

These are tools for money managers who are in ‘deploy mode’ and playing the odds.

Seed investors should not think of themselves as money managers first.

We’re people seekers and optimists above all else.

The largest outcomes in tech would have failed today’s automated filters at seed. Take the three $30B+ companies in FC’s Fund I:

+ The Trade Desk is adtech, a category that many investors won’t touch now and was invisible at the time. The founder had an English degree from BYU, not an MBA from Harvard.

+ Coupang was founded to become the Groupon of South Korea. Happily for all involved, it became the Amazon of South Korea instead.

+ Uber was a bizarre idea for an on-demand black car service, a Canadian founder pursuing a novel-at-the-time concept of “Full Stack Startups.”

But Jeff, Bom, and Garrett were unstoppable, and you figured this out when you took that first in-person meeting. All 3 of them had a burning passion for fixing a problem they saw in the world.

At the seed stage, it’s all about the people.

Many of FC’s highest-conviction investments are in products we suspect will pivot, because the people building them are so extraordinary and more likely than others to ‘see around corners’ in their industries. Exactly the types of startups that AI would screen out.

If you’re feeling overlooked right now: traction and customer love speak louder than resumes. Relentless determination is higher signal than a crowded term sheet. Early skepticism can mean unique insight, not impossibility, as long as you’re deeply in tune with your users and your market.

Venture has always demanded investors be non-consensus and right. Not just right.

Automated filters catch consensus. Human connections catch the rest. The partners who know the difference are looking harder than ever.

We’re looking for you.