Product-market fit has become harder to pin down in the AI era, even though startups are growing faster than ever. That was the starting point for a panel at TechCrunch's October conference, where Anu Bharadwaj of Iconiq, Rajeev Dham of Sapphire Ventures and Rahul Vohra, the founder of Superhuman Mail, discussed new rules for the concept. TechCrunch covered the session, and the advice is practical for anyone selling software right now.
The central message is about honesty with yourself and with investors. Customers are more willing to try new products, products are easier to copy, and revenue numbers can look better than the business underneath. Here is what the panel said about measuring real demand, which red flags to watch for and how to tell when to pivot.
Why pilots are not product-market fit
Bharadwaj said it has never been easier to collect a stack of pilots, because companies across the board are open to trying new things. The hard part, she said, is moving from pilot to production. A startup that counts every pilot as proof of demand can mistake curiosity for commitment, and the gap shows up later in retention.
That matches what many founders in the Bay Area describe when they sell to larger companies for the first time. A pilot gives you feedback and a reference, but a production contract gives you revenue you can plan around. If you are earlier in that journey, our post on founder-led sales and your first 10 customers covers how to get from first conversations to paying accounts.
How to report ARR without fooling yourself
The report says many founders pitch revenue metrics without saying which revenue they mean. Bharadwaj asked whether a number is true annual recurring revenue or a run rate, and she noted that there are many ways to blur the line. TechCrunch adds that chatter on X has focused on AI founders inflating ARR, for example by counting pilot customers who may use a product for only a couple of months as if the income will continue all year.
Her advice was to be upfront about the quality of your revenue, since honesty serves your long-term interest. It is a red flag, she said, when a few customers make up much of a company's revenue or when customers can easily switch to another vendor at a better price. Churn and retention depend heavily on those two conditions, so a founder who tracks them early can answer an investor's hardest question before it is asked.
Measuring engagement and dollar retention
Vohra recommended tracking customer engagement and dollar retention, particularly for consumer products, and finding out what counts as good, great and excellent for your category. At Superhuman, he said, the team tracked ARR and also an internal metric called "delighted ARR," which counted only revenue from customers who were deeply engaged with the product.
The idea is easy to borrow. Define what an engaged customer looks like in your product, count the revenue from those customers separately and watch whether that group grows faster than the rest. Vohra said that focus helped his team understand how sticky customers were and how long they were likely to stay, and he added that those users and companies are still customers today.
A simple test for when to pivot
For a green flag, Vohra suggested talking to customers about the value they find in the product and how they would feel if they could no longer use it, then being honest about the answers. He gave a rule of thumb. If your result is 15%, 14%, 13% or below, he said to consider a complete pivot, keeping the codebase and the team to build something new, while a result near 20% can be optimized toward 40%.
Bharadwaj said that at the seed stage it is fine to pivot, and that she does not look for product-market fit at that point. She looks instead for a plausible thesis and for a founder with the agility to adjust as things change. Dham added that he values founders who have made several pivots, because those moves show they can make hard decisions in a shifting market.
Vision is not the same as a business
Dham made a point that is easy to forget during a funding push, which is that a founder should not confuse having a vision with having a good business. A compelling story helps with recruiting and fundraising, but customers pay for a product that solves a problem they have today.
He also said product-market fit is never finished. Even at a billion dollars in revenue, he said, companies are re-earning their fit, which means constantly improving the product. If you are weighing a raise, our breakdown of Trajectory's $45M round shows how one early-stage AI company framed its story, and the guide to finding a cofounder in San Francisco helps if you need a partner to test these ideas with.
The same conference also hosted a session where investors from Peak XV, Index Ventures and Bessemer explained how to raise a Series A today, and one of them called a chart of cumulative ARR a major red flag. Reading both write-ups together gives a clear picture of what careful investors want to see before they commit to a round, and of how much they care about clean numbers.
Find your people
You will sharpen your metrics faster by comparing notes with other founders. Tribe organizes South Bay Renaissance to rebuild the South Bay and Peninsula startup scene in person, and its founder dinners and demo nights are a good place to test how you describe your traction. Browse upcoming gatherings on the Tribe Events calendar on Luma, and join the AI Trends Tribe to talk through ARR, retention and pivots with other builders.
