Partnering with Firetiger

AI was supposed to take some of the pain out of building software. It already writes a lot of code — about half of Google’s, by Sequoia’s telling — and Anthropic leans on agents close to 100 percent of the time. That’s the upside. The messier part is what happens after all that code piles up: it gets much harder to know when a product is quietly doing the wrong thing for customers.

Firetiger’s pitch is that observability shouldn’t be another thing engineers have to babysit. Instead of making teams manage events, dashboards and incidents in the usual way, the product runs agents around the clock. They look for anomalies, check whether those anomalies are real, figure out who is affected, and move to fix the issue before customers even notice anything is off.

That’s a more aggressive promise than the usual monitoring story. Firetiger isn’t just saying it will surface problems faster. It’s saying it can do the triage and the cleanup too, which is a very different job. If that holds up, it changes observability from a reporting layer into something much closer to an always-on operations crew.

Rustam and Achille are not trying to solve a fake problem from a slide deck. Rustam, as VP of Product at Cloudflare, worked on infrastructure that shipped to billions of users, and Achille handled observability and scaling problems at Twitch, Segment and Twilio. Sequoia says the product is already being used by everyone from the biggest companies in the world to the fastest-growing startups in Silicon Valley, and that it’s available fully self-service.

That combination matters. The people behind Firetiger have seen the failure modes up close, and they’re betting that modern software has outgrown the old dashboard-and-alert routine. Sequoia clearly likes the bet. It’s backing a tool that treats customer protection as a 24/7 agent problem, not an engineer-on-call problem.