My industry wants you to gut your sales team and replace it with AI. One of my customers plans to double theirs instead

EliseAI plans to double its sales team over the next 12 months. The company, which sells AI to housing and health care businesses and is a customer of my company, says its sales headcount has already grown 74% over the past year.

Every one of its salespeople uses AI, according to Ben Rey, its vice president of enterprise sales. EliseAI even builds some of its own tools because, he says, outside vendors can’t keep up. “We’ve tripled the productivity of our AEs, and many of them now close $5 million to $15 million per quarter each,” Rey says. “That productivity has us hiring faster than ever.”

At Attention, I build AI for sales teams, so this is a particularly interesting thing to hear. My industry has spent years selling the idea that software will let you stop paying salespeople. One vendor currently advertises, “Replace your $90k SDR with our AI Sales Agent.” The savings are easy to picture. A salary is a large expense, and a software subscription looks small beside it.

But a salesperson also brings in business. If AI helps that person win more of it, a company with room to grow has a reason to hire another. EliseAI is making that calculation. Its salespeople have become more productive, and it sees an opportunity to put more of them to work.

Those are company-reported results. They don’t tell us how much of the improvement came from AI rather than a strong product, a growing market, or the sales team itself. They do challenge the assumption behind the replacement pitch: that the best thing a company can do with a productivity gain is remove the person who achieved it.

What if instead of asking if AI can let employees go, companies started asking, how AI can make their next hire more likely to succeed?

For example, your best rep knows which questions to ask, when a buyer is being polite rather than serious, and which objection will kill the deal if it goes unanswered. You would love to hire ten more people like her. Instead, you hire people with potential and hope and pray they’re fast learners. Meanwhile, your star rep’s calendar is full, her manager is stretched thin, and the next batch of recruits needs coaching.

That skills gap gets expensive for the company when you have more customers to win than people equipped to win them. You can put the best rep’s calls in a training library, but the new hire still has to dig around to find the useful lesson and remember it during a conversation they’ve never had before.

AI can make some of that experience available when it is needed. Before a call, it can find how experienced colleagues handled a similar objection. Afterward, it can flag the question the buyer asked that nobody answered, or the promised follow-up that hasn’t gone out. The rep still has to understand the buyer and decide what to do. They have help with parts of the job that a more experienced colleague might do without thinking.

That changes the support a company can give its people. A manager can act on an unanswered question before the buyer has spent a week waiting. A new hire can prepare with the benefit of conversations they weren’t there for. A teammate taking over a deal can see what was promised. Fewer people have to spend their time recovering information the company already had.

In a study of 5,172 customer-support agents, economists Erik Brynjolfsson, Danielle Li, and Lindsey Raymond found that an AI assistant increased issues resolved per hour by 15% on average. Less experienced and lower-skilled agents gained the most, improving both speed and quality. Agents with two months on the job and AI assistance performed as well as agents with more than six months of experience without it.

The system had been trained on examples of successful work, including the behavior of top performers. It offered suggestions during customer conversations; agents could edit or ignore them. The strongest agents saw little benefit and small declines in quality, a result I’d take seriously before telling everyone to follow the same advice. The researchers couldn’t measure the effect on hiring. They did show that less experienced people could become effective much sooner.

This is what I mean by raising the quality floor. Your company shouldn’t need everyone to be superstar at every part of a job to deliver consistently good work. Better support can mean a new employee is able to be competent at their job more quickly, make fewer mistakes for their colleagues to fix, and give confidence to executives that adding people will really add to capacity.

For a sales leader, those kinds of improvements make a huge difference in planning. A new rep costs money to recruit, train, and pay before you know whether they will produce. If you can help that person become effective sooner and handle more opportunities well, the investment becomes easier to justify. Of course, you still need a product people want and enough profitable business to pursue. AI cannot supply those by filling in your CRM.

I don’t assume every employer will hire more. Stanford’s latest analysis of U.S. payroll data found employment for workers ages 22 to 25 in AI-exposed occupations was 19% below where it would have been, had it kept pace with their less-exposed peers. The gap operated mainly through reduced hiring. The authors call these descriptive findings, not causal estimates, but they are a serious warning. Young people can become better at a job with AI, but they still might never get the chance to prove it.

To be clear, I want an efficient company, too. At Attention, we try to automate work before hiring someone to do it. If there is a fixed amount of work and software can handle it reliably, maintaining a smaller team may make sense. But when there are customers still to win, improving the team can create an opportunity to expand it.

Employers should be asking whether new hires reach competence sooner, whether their work needs less repair, and whether customers get the answers they need. Those changes can make a hiring plan more ambitious. Measuring the number of emails sent tells us very little unless the people sending them are getting better at winning business.

EliseAI says its salespeople are so productive that it wants twice as many as it has today. I want to build tools that make companies like theirs make that decision easier: give the next person a better chance of succeeding, and give the company a better reason to hire them.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

#industry #gut #sales #team #replace #customers #plans #double