Every senior Amazon strategist I trust got good the same way: two or three years of grunt work. Bulk files at midnight. Search-term reports until their eyes bled. Case logs with Seller Support. A thousand small reps that nobody enjoyed and everybody needed.
AI just ate that layer. And the industry's response — confirmed in print this week — is to stop hiring the people who used to do it.
That's not a junior problem. Juniors are adaptable; they'll route somewhere. It's a you problem, three years out, when you go looking for a senior strategist to run your account and discover the industry quietly stopped manufacturing them in 2026.
What happened
On July 10, MarTech published "Who trains tomorrow's marketers if AI does the work?" by Milton Hwang, citing CMI's 2026 Career and Salary Outlook finding that one in three companies are reducing entry-level marketing hiring while simultaneously asking senior marketers to produce more with AI. It lands on top of BCG's "Mind the Marketing Gap" survey of 300 CMOs (June 15): 96% say AI is transforming marketing end-to-end, while only 8% actually run campaigns where AI agents operate autonomously — and 42% still use AI as a per-task assistant.
Why most people will read this wrong
The dumb takes come in two flavors, and they're both loud right now.
Flavor one: "AI is killing marketing careers." Doom content, infinite scroll of it. Three weeks ago I wrote that you should ignore the junior-role-death doomerism, and I stand by it — the juniors themselves will be fine in aggregate. They'll land in ops, in-house, or found something.
Flavor two — the one more of you are guilty of: "Great, leaner teams, lower retainers, efficiency." This is the CFO read, and it's correct for exactly four quarters.
The real signal is neither. Put the two studies side by side. Companies are cutting the entry layer (CMI) while running AI as a supervised assistant, not an autonomous workforce (BCG's 8%). That means the work still requires human judgment at every step — they've just stopped hiring and training the humans who were going to supply that judgment next. The industry is eating its seed corn and calling it a diet.
Here's the part that's specific to our world: Amazon expertise has never come from certifications. There is no degree in "why did my ACOS spike on Tuesday." Judgment in this industry is manufactured one way — reps on real accounts, with real money, under someone senior. The execution work AI just absorbed was the apprenticeship. Nobody designed a replacement. They just deleted the classroom and kept the tuition.
What actually changes for someone running $200K/month on Amazon
Walk the math on a typical agency pod. One senior strategist covering 4–6 accounts, one or two juniors underneath doing execution — bids, bulk ops, case logs, reporting. Junior fully loaded: $50–65K. The AI stack replacing their execution output: a few hundred dollars a month. That trade gets made every time, and in isolation it's rational.
Now run it forward. The senior on your account today has maybe 6–10 years of reps. The person who was supposed to replace them when they get promoted, poached, or burned out was supposed to be hired in 2025–2026 and spend three years getting good. One in three companies just didn't hire them. The 2029 supply of "actually senior" Amazon talent is being set right now, and it's being set low while demand holds flat or grows.
Three concrete consequences for your P&L:
Title inflation hits your account first. When real seniors get scarce, agencies don't leave seats empty — they promote the 18-month person and call them a senior strategist. Same line item on your invoice, a third of the reps behind the decisions. You will not get an email announcing this. You'll notice it in the quality of the "why" answers on your monthly call.
Key-person risk concentrates. Fewer real seniors means each one covers more accounts. When yours leaves — and their comp is about to get bid up, so they will — the drop-off to whoever inherits your account gets steeper every year. A senior departure that used to cost you a soft month now costs you a soft quarter.
The judgment premium reprices. Everything I've written since June points the same direction: execution is going to zero, judgment is the product. This is the supply-side confirmation. When the thing that's the product also stops being manufactured, its price goes one way. Expect senior-led boutique retainers to climb, and expect "AI-efficient" budget agencies to get cheaper while quietly having nobody senior in the building. The middle disappears.
What I'd do this week if I were you
-
Census your account team. Ask your agency, plainly: who touches my account, and how many years of hands-on-Amazon reps does each person have? Not titles — reps. If the answer is one gray-hair reviewing dashboards over 15 accounts and AI doing the rest, you're not paying for judgment, you're paying for a rubber stamp on model output.
-
Change your RFP question. "Do you use AI?" is dead — 96% will say yes and the answer contains zero information. Ask instead: "How do you train new senior strategists now that AI does the execution work?" Agencies with a real answer — structured review reps, teardown rotations, juniors auditing AI output against outcomes — are building a talent moat. Agencies that blink are three years from being a brand name wrapped around a model you could prompt yourself.
-
Ask the succession question. "If my strategist leaves, who takes over, and what have they already done on my account?" You want to hear that a second human has real context — not that "the team" (meaning the AI-plus-one-reviewer) has it covered.
-
If you run an agency: keep hiring juniors, change what they do. The old apprenticeship was execution reps. The new one is judgment reps — juniors review every AI-generated bid change, forecast, and creative brief against what actually happened 30 days later. Prediction, result, gap, lesson, logged. That compresses judgment formation from three years toward one, and in 2029 you'll be selling something your competitors structurally cannot hire.
-
If you have a brand-side junior doing ecom ops: don't cut them. You're watching agencies liquidate the exact asset that's about to get scarce. A junior who spends two years inside your account with AI leverage becomes the in-house operator you will otherwise be renting at 2029 prices.
What I'd ignore
The doom cycle. "AI killed marketing jobs" engagement-bait tells you nothing actionable. The jobs are moving, not vaporizing — the question is whether the training moves with them, and that's decided by operators, not LinkedIn posts.
The 80% comfort stat. BCG found 80% of CMOs investing "significantly" in AI training programs, and that number will get waved around as proof the pipeline is fine. Read what those programs teach: tool fluency. Prompting, platform features, workflow hygiene. Useful, and not the thing. Nobody's ACOS problem was ever solved by someone who knew the tools but had never watched a launch die. Tool training is not judgment training, and only one of them is scarce.
Upskilling certificates. A wave of "AI marketing certifications" is coming to fill this gap, and agencies will pin them to pitch decks. A certificate is proof someone watched videos. Reps on real accounts are proof someone made decisions and lived with them. Price accordingly.
Anyone selling this as a reason to fire your agency this quarter. The pipeline gap is a 2028–2029 problem you should position for now — not a reason to torch a working relationship in July. The move this week is asking better questions, not switching partners.
The uncomfortable summary: for twenty years, this industry got its senior talent for free — juniors paid the tuition in grunt work, and everyone upstream harvested the graduates. AI ended the grunt work, and the industry's first instinct was to pocket the savings and skip the school. The operators who build a new school own the next five years of talent. Everyone else gets to bid on the shortage.