There is a particular kind of institutional amnesia that afflicts businesses when a new technology makes the old way of doing things look embarrassingly inefficient. Agencies are experiencing it right now with the junior creative role — and the consequences, while not yet visible on any spreadsheet, are quietly building into something that will matter enormously.

The numbers are stark. According to Forrester, 15% of agency jobs will be eliminated in 2026 alone — a compression of what was forecast as an eight-year structural shift into a single calendar year. The culprit is not hard to identify: AI adoption across the industry has gone from 57% in 2023 to 86% by end of 2025, with the pattern in most agencies now being wholesale redeployment rather than incremental experimentation. The model, as one industry analyst bluntly summarised it, has flipped: "Historically, agencies made money through labour arbitrage — expensive creative directors overseeing cheap junior talent who did the volume work. Now it's small teams of senior talent working directly with AI assistants."

On the surface, this looks like progress. Copy generation is now the top AI application at 72% of agencies. Multimedia generation — images, video, audio — has jumped to 64%, reflecting how rapidly generative tools have matured. Superside reports delivering work up to five times faster at 40% lower cost than traditional agencies. Jellyfish, working with M&S, achieved 80% faster content delivery with a 30% reduction in unit costs using GenAI workflows. These are not marginal gains. They are structural transformations of the economics of creative production.

But here is what the efficiency metrics do not capture: the junior creative role was never just cheap labour. It was the industry's apprenticeship system. It was how you learned to write a decent brief before you wrote a brilliant one. It was how you developed an eye for what made a campaign sing before you were trusted to conceive one. The junior copywriter grinding through product description variants was, whether they knew it or not, internalising the grammar of commercial persuasion. The junior art director executing someone else's concept was developing the visual vocabulary they would later deploy with genuine originality. You cannot fast-track that process. You can only skip it — and discover, five years hence, that your pipeline of senior talent has an inconvenient gap in it.

The holding company behemoths are, characteristically, getting their infrastructure in order while the talent question idles in the background. At CES 2026, WPP unveiled Agent Hub — an internal marketplace of vetted AI agents embedded within WPP Open. Omnicom reframed its Omni platform as a unified operating system for strategy, execution and measurement. Havas outlined AVA, a centralised portal giving employees secure access to leading LLMs, due to roll out this spring. Stagwell introduced The Machine, an AI-powered marketing operating system designed to orchestrate workflows end-to-end. The vision is explicit: agencies as managed ecosystems of AI agents, built on proprietary data, wrapped in compliance. The humans who run those agents? Ideally, experienced ones.

Where, one might reasonably ask, will those experienced humans come from?

Creative Lives in Progress, which tracks early-career talent in the industry, has documented a marked deterioration in entry-level opportunities over the past eighteen months. The roles that survive — the ones deemed too creative, too relational, too judgment-dependent to automate — require precisely the kind of seasoned instinct that only accumulates through years of doing the work that AI has now absorbed. New roles are emerging, certainly: creative prompt engineers, diffusion engineers, AI quality stewards. But these are not entry points for people learning what good looks like. They are specialist functions for people who already know.

The industry that convinced itself it was solving an efficiency problem may have quietly created a succession crisis. The Cannes Lions winners of 2031 are, right now, the junior creatives who should be learning their craft on unglamorous client work — the product pages, the social content calendars, the below-the-line executions that nobody talks about but everybody needs. That work is vanishing into AI workflows with considerable speed and not inconsiderable glee.

There is, of course, a counter-argument, and it is not without merit: that AI tools will democratise creative skill-building, that the new juniors will learn differently, that the craft still transfers even if the medium has changed. Perhaps. But the agencies assuming that future without actively designing for it — without building the mentorship structures, the deliberate AI-assisted learning pathways, the protected spaces for craft development — are betting rather heavily on optimism. The efficiency gains are booked. The apprenticeship replacement plan remains, for most agencies, a blank slide in a deck that nobody has written yet.