These 20 and 22-year-olds raised $5 million from YC, a common catalyst for studying online behavior using Vision AI.

Amogh Chaturvedi runs around a bit of sleep, but there is a lot of certainty at 6am, he is a gloggy, apologising for the rescheduling, and is still upset by recent horrors, including family and an electric scooter.

Within minutes, however, the 20-year-old Stanford Dropout had focused, as he and his co-founder sold one startup at the age of 19, landed in Y Combinator, raising $5 million for the next company, the human behavior.

Human behavior, launched just a few months ago, bets that Vision AI can do what analytical tools like Mixpanel and Posthog struggle with.

Instead of relying on manually tagged events and clickstream data, Human Behavior claims that AI generates replays and insights of real user sessions, answering product teams’ most pressing questions without hours of instrumentation code.

The four-month YC startup has shut down its $5 million seed round (which is becoming the norm for today’s YC companies) in just two days, with supporters such as General Catalyst, Paul Graham, Vercel Ventures and Y Combinator.

“We could have done a financial engineering game because we got more offers with higher ratings, but we didn’t want that,” the CEO said.

Chaturvedi met his co-founders Skylar Zi and Chirag Kawediya at the Hacker House, which he organized in 2023 as an excuse he built up and lived with friends after his freshman year at Stanford.

Their first startup, the fabric, was the e-commerce accounting tool they bootstraped. Like Chaturvedi, Ji dropped out of college (leave Berkeley), and Kawediya graduated.

YC was initially skeptical of the potential of the fabric market, but the team was hospitalized this year in an accelerator spring batch, assuming it would eventually pivot, Chaturvevi says. They did so almost immediately after talking to all the customers and asking about other issues they faced.

The feedback was consistent. The fabrics were able to show what products they were selling, but the customers wanted to know why. It answers the necessary analysis not only with accounting reports but also with behavioral data.

With this new direction, the team sold the fabric to Employer.com with six figures, bought benches and sold to the same company that carried out all human actions.

Kawediya explains that companies using traditional analytics often require engineers to set up event trackers for each button.

For fast-moving startups, that’s far from ideal. “Even if you get that data, you can improve it because you’re stuck with the bigger questions about how users actually interact with the product,” he says.

Although session replays are not new, until recently, computer vision models were not accurate enough to analyze at scale. Now, human behavior does so to summarise and segment thousands of hours of footage. “Do you spend hours writing code to track clicks when you can watch videos?” adds.

Today, human behavior customers (mostly fast Series A and B startups) get daily summary emails highlighting the features used, which bugs are displayed and the user churned. Since its launch four months ago, Chaturvedi says the company has grown 20% a month.

The founder calls the session. Nowadays, human behavior helps teams understand users and squash bugs. Over time, the same dataset was able to power automated QAs and embed IT support. Their ambition is to turn human behavior into a data dog for session replays.

Building new technology from the ground up is a way that founders believe they will take on more established players like Mixpanel and Posthog. “For some of these companies, it may be difficult to replicate what we have because their construction can’t support the shift without having to start over,” Chaturvedi said.