By Urban Gavelin, , AI & Sales Capacity
Your Next Salesperson Might Already Work for You
Everyone is watching for layoffs. How many jobs will AI destroy? Which roles are most at risk? When will the big wave hit? I think we are looking at the wrong thing.
The question nobody puts on a slide
I had lunch with a salesperson recently who works closely with developers. We started talking about AI and productivity, and landed on the same observation from two completely different directions.
The first big AI effect may not be that people get let go. It may be that the next person never gets hired at all. A skilled developer with the right AI tools can now produce significantly more than the same developer could a couple of years ago. That does not have to end with, 'Thanks for your time, we are letting you go because of AI.' It is more likely to end with a much quieter sentence in a leadership meeting: 'You know what, maybe we do not need to hire that person after all.'
That is exactly why the shift is so easy to miss. Nobody writes an article about a hire that never happened. Nobody holds a press conference. It just... does not occur.
Do the math on what you already have
When I look at sales, this gets even more interesting. Sales organizations have long worked from a fairly simple equation: if we want to sell more, we need more salespeople. But what if that equation is starting to break down?
Take a ten person sales team. A large chunk of the work week can disappear into things that are not really selling: moving information between CRM, spreadsheets and email, writing meeting notes, hunting for information, updating systems. Somewhere in there sits a highly trained salesperson acting as a human integration layer between two systems that refuse to talk to each other. It would almost be funny, if it were not so expensive.
Let us say roughly 70 percent of the time goes to administration and other work around the actual selling. That means you are effectively paying ten people to get about three full-time equivalents of real sales work. The traditional answer is to hire person number eleven: new salary, new onboarding, months before they are fully productive, and then you drop them into the same broken system as everyone else, where they too will lose most of their week to admin. It is a bit like discovering the bathtub is leaking and solving it by buying a bigger tap.
There is another way to think about it. Instead of asking 'Who should we hire?' , ask 'How much capacity can we buy back from the people we already have?' Free up 20 percent of the time across ten salespeople and you have effectively created two full-time roles, without two recruitments and without two onboarding processes. You may have just found your next two salespeople inside the organization. They were simply stuck in the CRM.
Why doesn't everyone do this?
Because it is uncomfortable. You actually have to look at how the work gets done, not how the process diagram in the slide deck says it gets done, but how it really happens. Where do we type the same information twice? Which handoffs create extra work? What are people doing manually that a system should be doing for them? And perhaps the most important question: what are we doing today that should not be done at all?
It is far more fun to talk AI strategy, ideally with a slide that says 'Transformation' next to a futuristic blue robot. But transformation rarely starts there. It usually starts with someone sitting next to a salesperson and asking, 'Why do you do that?' 'Because the CRM needs the information.' 'But the information is already in the email, isn't it?' '...yes.' And there, suddenly, is a rather interesting AI opportunity.
There is also a more human reason we default to hiring. A new hire looks like momentum. 'We are growing. We are bringing on five new salespeople.' That sounds like something is happening. Removing ten hours of admin per week looks like almost nothing. Nobody pops champagne because the CRM updated itself automatically. But a few months later the results show up: more customer calls, faster follow-up, more time in genuine dialogue with customers. More human time spent where human time actually matters. That is why I think one of the most common unspoken beliefs in B2B needs to be challenged: that more revenue requires more headcount. Sometimes it does. But far from always. In many organizations, growth is first and foremost an architecture question, not a recruitment question.
Do this on Monday
Ask three people on your team to write down everything they did the previous workday, in fifteen minute blocks. Not what they usually do. What they actually did. Then mark every block as customer time or not customer time . Preparing for a customer meeting is customer time. Manually re-typing information into the CRM that already exists in the meeting notes is not.
Add it up. That is your real capacity. Then take the largest non customer time category and ask a single question: why is a human doing this? If the answer is 'because the systems do not talk to each other,' 'because someone has to compile it,' or 'because we have always done it this way,' you may well have found your next hire. Just without the hiring.
So here is the question worth sitting with: if you could recover the equivalent of two full-time people from the team you already have, would you still want to hire the next person? The first big AI effect may not show up in statistics about who lost their job. It may show up in all the jobs that never needed to be created. And that is a much quieter revolution.
Frequently asked questions
Is this article arguing that AI will not cause job losses?
Not exactly. Urban's point is that the more common and less visible effect is roles that simply never get created, because existing employees regain enough capacity to absorb the extra work. That shift rarely makes headlines, but it can be just as significant as visible layoffs.
How do we start identifying hidden capacity in a sales team?
Start by having a few salespeople log their actual activity in fifteen minute blocks for one working day, then classify each block as customer time or not. Whatever category dominates the non customer time is usually the first place to investigate for automation or process fixes.