A Y Combinator startup has launched an AI platform that automatically fixes product issues by watching user session replays and taking action without human intervention.

Human Behavior's system captures user events, errors, and session recordings, then deploys AI agents to analyze every replay for problems like rage clicks, broken buttons, and user abandonment points.

The platform goes beyond traditional analytics by automatically responding to detected issues. When the AI spots a problem, it emails affected customers, updates project management tools like Linear, opens pull requests with video evidence, and tracks whether fixes actually resolve the underlying issues.

"Dashboards are where insights go to die," said founder Amogh Chaturvedi in the Product Hunt launch post. "Every product team is sitting on thousands of hours of session recordings and charts nobody has time to look at."

Dogfooding Results Show Promise

The company has been testing the system on itself with notable results. This month, Human Behavior's AI agents analyzed 15,000 sessions of people using their own product and automatically opened 16 pull requests in their codebase.

The fixes addressed dead buttons, silent errors, and layout shifts — none of which had been reported as bugs by users. The autonomous system identified and proposed solutions for issues that would have otherwise gone unnoticed.

Human Behavior started as an AI accounting platform before pivoting multiple times over 1.5 years of development. The current product emerged from an internal tool the team built to monitor their own user sessions.

The platform integrates with existing workflows through Slack and SMS notifications, eliminating the need for teams to monitor separate dashboards. All insights and actions flow through communication channels teams already use.

The startup is currently taking calls with potential customers through their website at humanbehavior.co. The company positions itself as competing with traditional product analytics tools like PostHog by adding autonomous remediation capabilities.

Human Behavior plans to expand the system's ability to detect and automatically resolve more complex user experience issues as the AI models improve.