AI Took Over the Junior Marketing Work. It Also Took the Only Way to Train a Senior.

In September 2025, the labor-analytics firm Revelio Labs reported that US entry-level job postings had fallen about 35% since January 2023, more than 100,000 fewer openings a month. CNBC covered the data, noting some tech and data roles were down as much as 67%. Big tech’s new-graduate hiring, per the venture firm SignalFire, has dropped more than 50% from pre-2023 levels.

Those numbers describe a knowledge-work squeeze in general. Marketing sits near the center of it.

The junior marketing job was never only about the output. It was the tuition.

In July 2025, Microsoft researchers published “Working with AI,” an analysis of 200,000 anonymized Copilot conversations mapped against occupations. Writers, public relations specialists, and market research analysts landed among the most exposed roles, and the study found higher AI applicability for jobs that require a bachelor’s degree than for those that do not. Kiran Tomlinson, a senior researcher at Microsoft and one of the authors, told Fortune the finding is about tasks, not headcount:

“Our research shows that AI supports many tasks, particularly those involving research, writing, and communication, but does not indicate it can fully perform any single occupation.”

Research, writing, communication. That is the entire job description of a marketing coordinator’s first year.

The marketing-specific data points the same direction. In Wynter’s survey of B2B SaaS marketing leaders, content and copywriting topped the list of AI-exposed functions at 60%, with junior and entry-level roles named by another fifth of respondents. We covered how that reallocation is already happening quietly, through attrition rather than layoffs, and how job postings keep rewriting what a marketer is asked to do.

Agencies feel it first. MarTech reported that 57% of agencies have slowed or paused entry-level hiring as AI absorbs the execution work juniors used to own. Call it compression from below: the senior marketers stay, the rungs beneath them thin out, and the industry quietly stops manufacturing the people who replace the seniors when they leave. Harvard economists Seyed Hosseini and Guy Lichtinger, studying résumé and job-posting records from more than 280,000 US firms, found junior roles shrinking sharply at companies adopting AI from the first quarter of 2023, driven by reduced hiring rather than firing. They described it as an erosion of the “bottom rungs” of internal career ladders.

The mechanism is not mysterious. The work AI does best is the work juniors used to learn on. Pull the campaign’s numbers. Build the prospect list. Draft the first version of the email. Resize the banner. Chase the status update. A senior marketer with a model clears that queue before lunch, and the output reads as a win on this quarter’s budget. The cost arrives later, when there is no one who spent three years doing that work badly, then adequately, then well, and is ready to run the team.

This is the part the savings math skips. A junior seat was always two purchases bundled into one line item: the work that seat produced, and the marketer that seat produced. AI commoditized the first. It did nothing for the second. Cut the role and the quarterly number improves while the five-year bench empties, and no dashboard shows the second effect until the senior you needed is not there.

Some agencies are trying to rebuild the training ground instead of removing it. Jennifer Spire, CEO of the Minneapolis agency Preston Spire, has redefined junior roles around guiding and auditing AI output rather than producing first drafts, with junior staff frequently teaching senior colleagues how to use the tools, MarTech reported. That is one workable answer. It also requires a leader to decide the apprentice seat is worth paying for when a software subscription is visibly cheaper, a decision that runs straight into the cost pressure squeezing marketing leadership.

For anyone running a team, the entry-level decision has quietly split in two. The old question was whether a coordinator’s output justified the salary. The new question is whether the training that seat provides justifies the salary, because the output is already cheap. Most budgets only price the first half, which is how this shows up across AI and marketing jobs: as a saving today and a shortage later.

Three moves follow from the data. Treat a junior hire as a multi-year training investment with a return that never lands inside the quarter you approve it. Rebuild the first-year role around the judgment work models still cannot do: reviewing output, talking to customers, owning one small project end to end. And count the people on your team who could run it in five years. If the honest answer is no one you are hiring now, the AI savings are a loan taken against your own bench.

The tools got good at the apprentice’s tasks. Nobody has replaced the apprenticeship.