John Margerison on the new class of employee: AI managers

A new class of employee, the "AI manager", will emerge within the next two to three years as businesses restructure their workforces around artificial intelligence, according to John Margerison, CEO and founder of communication intelligence firm XFactorAI. This new class of employee will make the cut in the AI adoption overhaul: those who can prove their worth and those who can manage AI systems.

Staggering job layoffs are taking place across the globe: contract workers are finding themselves out of work, while entry-level positions are becoming harder to get. In January of this year, Salesforce cut over 1,000 jobs across product management, marketing, and data analytics, while Oracle cut 21,000 jobs over the past 12 months, citing the adoption of AI technologies.

The need for upskilling is here now, according to Margerison. And it's not going to be learning how to use an LLM or an agentic system; it will be learning how to craft quality prompts that will reign supreme as AI Managers prove their worth.

Margerison added that the spread of AI-driven working practices will not be determined by geography but by individual curiosity, specifically the quality of the prompts employees ask and how much thought they put into using the technology.

"You've gone from a normal bicycle to then now we've got 10 speed bicycles. But if you don't know how to ride a 10 speed bike, it's not going to work out really well. Some people are getting very technical as to go - 'great. I've got now all these new tools, and I can do potentially 10 times the amount of work what somebody who didn't have the tools can do,'" he said. "I think if you're looking at how to create more value in the marketplace than you're an employee saying, 'how do I get a keep my job, or how do I get a pay rise, or how do I dramatically increase my value to any company that wants me?'"

He added that the model's quality needs to be built into the infrastructure. Vibe-coded models that do the job one day can break down or become easily infiltrated the next. 

Margerison described himself as a strong supporter of keeping people involved in AI-driven decisions, while also backing automation where it is appropriate. He said the degree of oversight required should scale with the potential impact of a given application, contrasting low-stakes uses, such as managing a calendar, with higher-risk uses in regulated sectors like healthcare, where errors could directly affect patients.

Margerison said the obligation to reskill the workforce lies with both employers and employees. Companies benefit most from investing in staff who already understand the business and are willing to upgrade their skills, which they call a natural fit for both sides. Workers who resist adopting new AI tools, however, risk being overtaken, since the technology is advancing quickly and delivering real results - leaving companies to bring in someone with a different, more current skill set. As with any new hire, he said, that trust has to be earned gradually rather than granted all at once.

"You've got a new employee who started at this workplace. Do you give them access to all records? Do you let them make all the decisions that they want to make? What do you say? 'Hold on, sec, you just started with us. How about you do this part of this work, and we'll give you access to these records and this information.' When they've been there for six months or a year, you say, 'I want you to do some extra roles, you've shown capability, and you've shown we can trust you."