Amid growing tension over workplace AI, agencies have a window to reimagine staffing

At many agencies, fluency means using and developing AI agents to automate planning or media activation work. A Digiday survey published in May found that just 9% of media agencies were not pursuing agentic AI solutions of some kind.

That automation has saved time for individual workers, but few companies have successfully distributed AI gains across teams or departments. According to a recent estimate by Apollo Partners, while tech companies have improved average profit margins, and revenue-per-worker have risen among U.S. businesses since the introduction of generative AI in 2022, neither measure has improved among companies outside that sector.

That might change as the over-use of AI in corporate settings begins to make working life more tedious and more stressful — and brings tensions over individual usage of AI to the fore.

Agency attendees, speaking under Chatham House Rules at the Summit’s townhall session, said that colleagues both junior and senior are using tech as an all-purpose corner cutter — speeding up some processes but, in the same act, skipping over opportunities for critical thinking.

“I’ve been surprised by how much AI content I’m seeing come from vice presidents and C-suites,” said one attendee. “I don’t think they want to take the time.”

With that problem in mind — and the desire to capture and institutionalize AI productivity gains beyond experimental programs — some agencies are making space for staff education. At Monks, Wilensky said training of junior staff (and re-training of senior colleagues) held part of the answer. 

“If we don’t provide that, things can go off the rails,” said Wilensky.

Nadja Bellan-White, group CEO of M&C Saatchi North America, said her agency had carved out an hour each week on employee timesheets that’s dedicated to training. “A learning agenda has to be operationalized throughout your organization,” she said. “You have to have a culture of always learning, always pushing forward.”

Monks was also beginning to review hiring for its early-career roles, Wilensky added. “We’re not hiring less for junior talent. But it’s [being] more strategic about who you hire and what tasks you want junior talent to do,” she said. “I don’t think that junior team members are going to be sitting [and] piecing spreadsheets and recording spreadsheets all day… that work is taken off of their plates.”

Meanwhile, access to tools is limited for some job roles, with developers or engineers getting access to more powerful (and more expensive) AI models — a measure principally designed to control compute costs, but which can serve to limit exposure to AI tools where they’re unhelpful.

Such measures represent early efforts to reconcile traditional ad agency staffing models with AI. 

There are signs that AI is provoking some marketers to rethink their jobs. Freelance programmatic and strategy consultants have become avid users of the technology, while the Wall Street Journal recently reported a rise in companies employing marketing ‘engineers’, whose expertise was augmented by AI tooling.

Agencies, however, are only just now beginning to seriously consider how teams might be organised differently. That isn’t itself unusual — it took roughly forty years for factory owners to take advantage of the productivity gains offered by electric light.

Agencies won’t take as long as that to begin dialing in AI gains into team structures. Wilensky suggested that senior talent would be required to inject originality and a willingness to push the envelope into agency outputs.

Related Insights

Business of AI

‘There’s no room for purists’: Generative AI is altering the agency junior talent search Read More

“AI cannot take risks. That’s always going to be our job. AI is not creative. That’s always going to be our job. So no matter what level you are in your career, there’s opportunity to be strategic, to be creative, to take risks, to test,” she said.

Early career options, she added, may instead become focused on quality assurance or maintaining the fences agencies build to prevent AI agents from going rogue.

“We don’t have this role today, but there will be a fulltime job of AI checkers. That’s my premonition,” said Wilensky. “That’s going to be a career that our kids will do when they are older – there’ll be AI checkers in the same way that people in journalism used to be like fact checkers.”