I went into day one of TechCrunch Disrupt 2026 with a different mindset than I’ve had at past events. I wasn’t trying to take in everything. I was looking through all the noise and activity for the conversations that caught my interest. When something grabs me, it’s usually tied to something I’m already working through, even if I can’t name it yet.
A panel called “PMF Red Flags: How To Tell If You Really Have It” was one of those. Anu Bharadwaj from ICONIQ, Rajeev Dham from Sapphire Ventures, and Rahul Vohra from Superhuman were on stage, with Dominic-Madori Davis from TechCrunch moderating. I stood in the back and listened.
Product-market fit used to sound like a finish line. You find it, and you’re done. That idea doesn’t hold up anymore. With AI, products can change every few weeks. What customers expect changes just as fast. So the real question isn’t whether you found fit once. It’s whether you can keep finding it.
Three things stuck with me, one from each of them.
Lesson 1Founder-market fit
Anu talked about founder-market fit. She said that at seed, she doesn’t look for product-market fit. She looks at the founder. Does this person have a thesis that holds up? And can they adjust when things change? What I heard was ability and agility. Can you do the work, and can you move when the evidence moves?
The red flag is the opposite: a founder so sure their first idea is right that they won’t change course, even when the evidence says they should. I’ve seen this with founders I’ve worked with. They hit the edge of an iceberg and think it’s the whole thing. They found something real, but the bigger opportunity was right next to it.
The same thing happens inside companies buying AI. A team picks a tool, rolls it out, and holds on to the first plan even when the numbers say it isn’t working. Being sure feels like strength. But the real strength is noticing early and changing on purpose. Changing course isn’t the same as giving up. It’s what you do when you take the evidence seriously.
Lesson 2How you handle compression
What I took from Rajeev was how founders respond under compression. What tells him the most is how a founder handled the hard moments: stripping the product down, rebuilding it, even walking away from customers who weren’t the right fit. Those calls say more about a founder than whether they stuck to the original plan.
That one got personal for me. I’ve had big wins on this build, and I’ve had big losses. I’ve learned something real from both. Sitting there, I realized the losses aren’t something to hide. They’re part of the proof. What did I set out to do? What happened? What did I change because of it? That’s the win in the loss.
A simple way to show this is to write down three things every time you change course: what you believed, what you saw, and what you did next. Over time, that becomes a record of how you think. It’s also a good check. If you can’t explain why you changed something, it’s worth asking whether you should have.
For teams, compression often shows up as an AI pilot that isn’t going as planned. The easy move is to push harder, or to let it quietly fade. The better move is to stop, strip it down to the one job that matters most, and test just that. Walking away from a use case that doesn’t fit isn’t failure. It’s the same kind of call Rajeev described, made inside a company.
Lesson 3Name your own unit of value
Rahul talked about founders defining their own unit of value. You decide what real value looks like for your product, then show it growing. At Superhuman, he said, that was emails sent, not emails read. Reading is busy work. Sending is the job getting done.
That stuck because it’s a question I keep coming back to. It’s easy to say customers love what you built. It’s harder to name the exact thing they get from it. Naming it yourself shows you understand what you’re really building.
This works for teams using AI, too. Counting how many people logged in to a new tool is a lot like counting emails read. It looks busy, but it doesn’t tell you much. The better question is what got done because of the tool. A deal that closed faster. A report that used to take a day and now takes an hour. A customer who got an answer the same afternoon. Name that, count it, and you’ll know if the tool is earning its place.
TakeawaysWhat I’m taking with me
Put together, the three lessons point to one idea. Fit isn’t something you find once. It’s something you keep proving. You prove it with a thesis you can defend, with the calls you make when things get hard, and with a number that shows real value growing.
Here are the three questions I’m taking back to my own work:
- What do I believe, and what would change my mind?
- What did I change in the last month, and why?
- What is the one thing my customers get that I can count?
I also walked out with something I didn’t expect to find: hope. Part of me had been worried I’d missed my window. But if what matters now is how a founder thinks, learns, and adapts, then AI gives me the tools to show that, without being gated by pedigree. When everyone has the same tools, the person using them makes the difference.