According to Fortune, Cursor CEO Michael Truell revealed at the Brainstorm AI conference that the AI coding startup has automated roughly 80% of its own internal customer support tickets using its technology. The company, valued at $29.3 billion, also uses an internal AI system that lets employees ask any question about the company and get an AI-generated answer. Truell said Cursor crossed $1 billion in annualized revenue last month and now has over 300 employees. He cited a recent University of Chicago study showing teams using Cursor’s AI assistant merged 39% more pull requests than non-users. Interestingly, the research indicated senior developers were more effective with the AI tools than juniors, a finding that surprised Truell.
The Ultimate Dogfooding Test
Here’s the thing: every tech company talks about using its own product. But Cursor is taking it to an extreme, and that’s genuinely fascinating. They’re not just letting engineers use the coding assistant. They’ve built an entire internal AI help desk that handles the mundane stuff—password resets, access requests, basic IT questions. Freeing up human support for the complex, weird problems is a no-brainer. But the internal “ask anything” AI comms system is the real eyebrow-raiser. That’s a bold move. It means they’ve presumably connected their AI to all sorts of internal data—HR policies, project docs, sales figures. The potential for leaks or bad answers is huge, so their confidence there speaks volumes.
The Senior Developer Surprise
Now, the most counterintuitive part of this whole story is who benefits. Truell admitted he was surprised that academic research pointed to senior engineers getting more value from AI coding tools like Cursor than juniors. This flies in the face of the common narrative that AI is a great equalizer, boosting junior productivity the most. But think about it. A senior dev knows the architecture. They know what “good” looks like. They can craft a precise prompt, evaluate the AI’s output critically, and integrate it seamlessly. A junior might get bogged down debugging AI hallucinations. This aligns with that METR study from July 2025 which found experienced devs actually took longer with AI, perhaps because they were doing more rigorous review. It’s not about speed; it’s about leverage and quality.
The Enterprise Reality Check
And this is where Cursor’s internal experiment hits the real world. The article notes that big companies face “data silos” and “technical sprawl.” That’s the rub. Cursor, as a young startup, can wire its AI into everything from day one. A 30-year-old Fortune 500 company? Forget it. Their data is locked in a dozen legacy systems. This is why Cursor’s reported productivity gains and its own internal success might be hard to replicate everywhere. It requires that “dedicated technical expertise” to tailor the models—exactly what Cursor is doing with its forward-deployed engineers building custom tooling for sales and ops. They’re essentially building the blueprint they’ll eventually sell.
Beyond the Hype
So what’s the real takeaway? Cursor is proving that AI’s biggest impact might be as an internal operations and force-multiplier tool, not just a coding sidekick. The fact that they’re using it to run their own business is a powerful case study. But it also shows that successful AI integration isn’t just about buying a license. It’s about deep customization and having the skilled people who know how to use it effectively. The senior dev data suggests that throwing AI at a problem without expertise might even slow you down. Basically, AI isn’t a magic wand for incompetence. It’s a power tool for the already capable.
