Jason Lemkin, known to some as the godfather of SaaS, says now is the time to push the boundaries of AI in the workplace.
In practice, Lemkin, the founder of SaaStr, the world’s largest business-to-business community of founders, recently said on Lenny’s podcast that this means he will stop hiring humans in his sales department.
Instead, SaaStr will fully on the agents, which are commonly defined as virtual assistants capable of performing tasks autonomously. They break down problems, outline plans, and take action without prompting from a user.
He said the company now has 20 AI agents that automate tasks once handled by a team of 10 sales development representatives and account executives.
This shift from an entirely human workforce to a agent-based workforce was quick.
As of May, SaaStr had just one AI agent in production that it used for various digital tasks, Lemkin said. That month, however, at the SaaStr Annual – its annual gathering of more than 10,000 founders, executives and venture capitalists – two of its well-paid sales representatives abruptly resigned.
Lemkin said he turned to his head of AI and said, “We are done with hiring humans in sales. We’re going to push the boundaries with agents. »
Lemkin’s calculation was that it simply wasn’t worth the cost of hiring another junior sales rep for a $150,000-a-year position who would eventually quit, when he could use a loyal AI agent instead.
Amelia Lerutte, SaaStr’s director of AI, told Business Insider via email that in June the company began increasing the number of agents in production.
“We only had one non-essential agent at the time at Delphi, but we didn’t delve 2-20+ until early June,” she said. “It was a conscious choice after their departure to reallocate some (but not all) of the staffing costs to the agents.”
At SaaStr’s office, the 10 desks that once belonged to humans on the marketing team now have agent names, like “Quali for Qualified,” “Arty for Artisan” and “Repli for Replit,” Lemkin said.
Lemkin said SaaStr trains its agents on its best humans.
“Build an agent with your best person and your best story, and then that agent can start to become a version of your best salesperson,” he said.
SaaStr’s process is similar to that of Vercel, the cloud-based platform for developerstrained a sales agent under their top performer for six weeks by documenting every step of their work, then creating an agent to mimic their process.
Many companies are experimenting with AI agents, but risks remain. One of the biggest is the threat of data leaks and cybercrime.
“AI agents, to benefit from their full functionality and to be able to access applications, often need access to the operating system or operating system level of the device you’re running them on,” Harry Farmer, a senior fellow at the Ada Lovelace Institute, recently told Wired.
All of this access creates more potential points of attack for cybercriminals.
Security threats aside, Lemkin said the net productivity of agents is about the same as that of humans. However, he added, agents are more efficient and can scale, just like software.