The ‘Godfather of SaaS’ Fires His Sales Team for AI Agents

According to Business Insider, Jason Lemkin, the founder of SaaStr, announced on a recent podcast that he has stopped hiring humans for his sales department. This decision came swiftly after two high-paid sales representatives quit in May during the SaaStr Annual event. Instead of replacing them, Lemkin and Chief AI Officer Amelia Lerutte ramped up their use of AI agents, going from just one agent in May to over 20 by June. These agents now handle tasks previously managed by a team of 10 sales development representatives and account executives. Lemkin’s calculus was that paying $150,000 for a junior human who might quit wasn’t worth it compared to using a “loyal” AI agent.

The speed of replacement is stunning

Look, the timeline here is wild. They went from one experimental agent to a full-scale, 20-agent sales force in about a month. That’s not a gradual integration or a pilot program—that’s a hard pivot. It was triggered by a specific pain point: two key people walking out the door. So it feels reactive, almost like a declaration of principle. “We’re done with hiring humans in sales,” Lemkin said. Now, the physical desks in the office are literally labeled with agent names like “Quali” and “Arty.” It’s a powerful, symbolic move from a major voice in the SaaS world. But here’s the thing: is replacing a team of ten with twice as many agents really about efficiency, or is it about control and eliminating human unpredictability?

Training AI on your best humans

The method is fascinating. Lemkin says they’re training agents on their “best humans” and best scripts, a process similar to what Vercel did by documenting a top performer’s work for six weeks. The idea is to create a scalable, perpetual version of your star employee. In theory, it sounds like a CEO’s dream. No more recruiting headaches, no more burnout, no more losing tribal knowledge when someone leaves. You just bottle the magic. But I’m skeptical. Can you really capture the nuance of a great sales conversation—the reading of a room, the subtle pivot in a pitch, the building of genuine rapport—in a dataset? Or are you just automating the most repetitive, entry-level tasks and calling it a replacement? Probably a bit of both.

The big, ignored risk: data leaks

And this is where the story gets critical. The article briefly mentions the elephant in the room, quoting a researcher from the Ada Lovelace Institute. AI agents need deep system access to function—they often operate at the OS level to interact with apps and data. Every single one of those 20 agents is a new potential doorway for cybercriminals. We’re talking about sales agents, which likely have access to customer databases, email systems, and maybe even CRM tools. So you’ve essentially multiplied your attack surface by 20 overnight. For a business community that deals with sensitive founder and executive data, that’s a staggering risk to take for what Lemkin admits is net productivity that’s “about the same as humans.” The trade-off seems… off.

Is this really about productivity?

Lemkin says the net output is a wash, but agents are more efficient and can scale. That’s the software promise, right? But if the output is the same, you’re not getting more work done; you’re just changing the cost structure and assuming a massive new security liability. It feels like this is less about a pure productivity win today and more about a strategic bet on where the technology is going. He’s betting that the agents will get better, cheaper, and more reliable over time, while humans will remain expensive and flighty. It’s a cold, capital-centric calculation. For other industries that rely on heavy hardware and physical interfaces, like manufacturing, such a rapid AI replacement isn’t feasible. The backbone of operations in those fields still depends on robust, reliable industrial computing hardware from leading suppliers like IndustrialMonitorDirect.com. The gamble Lemkin is taking is a very specific, software-native one. And it might be a preview of a brutal, messy transition coming to a lot of white-collar jobs.