Entry Hiring Stays a Fifth Below Trend Through 2029

We forecast that entry hiring in law, finance, and software stays about a fifth below trend through 2029 and that most large firms stop funding junior training.

Key points

• Entry hiring in AI-exposed knowledge work will likely stay at least 20 percent below trend through 2029.
• Fewer than a third of large law and finance firms will fund junior training as a budgeted line.
• Fixed fees remove the payer for junior hours even where agents still fail the work.
• A visible senior shortage in finance, law, and software is the likely cost by 2030.
• We put about 65 percent on the full call.

Our forecast

Through the end of 2029, every Canaries revision will show employment of workers aged 22 to 25 in AI-exposed occupations at least 20 percent below trend. Fewer than a third of the 50 largest US law firms and the largest finance employers will fund junior training as a budgeted line. We put about 65 percent on both holding.

The rung narrows without mass layoffs. Firms stop hiring at the bottom.

Why we think so

Our signal scan found agents reaching the junior workload before firms rebuild training. The best agent completes five-hour tasks half the time¹. That horizon doubles every 131 days¹. Employment of young workers in AI-exposed occupations already sits 19 percent below trend². The prior estimate was 13 to 16 percent². The share of CEOs planning to move away from junior roles rose from 17 to 43 percent in a year³.

Three Open Questions on Agents and the Junior Workload showed that agents still fail most real projects. The top agent finishes about 16 percent of paid freelance work⁴. Capability alone would argue for a slower cut.

Pricing is the second force, and it moves on its own schedule. Citigroup asks bidding firms to show AI savings⁵. Morgan Stanley plans more fixed fees⁵. Zendesk bills 1.50 dollars per automated resolution⁶. A fixed fee removes the junior hours that paid for training⁵. Training then comes out of partner profit. At one Thomson Reuters forum, nearly every table of law firm leaders protected the AI platform and trimmed associate pay⁷. Either force alone would narrow the rung. Together they make the cut hard to reverse.

Warehouses, Agent Files, and Compute Bills Reshape the Entry Rung found the same shape in fulfillment centers. Headcount there fell for years through reduced hiring while output rose. Cheap capability spreads the pattern. Open-weight models trail the closed frontier by about four months⁸. Small firms will gain junior-grade agents soon after large ones.

A minority will fund training like simulator hours⁷. Ropes and Gray credits a fifth of first-year billable hours to AI learning⁹. Agents Take the First-Year Work and Thin the Senior Bench traced where thin funding leads. We see no sign that most firms will follow before 2029. Policy will lag. Disclosure rules cover layoffs¹⁰. A firm that stops hiring juniors files nothing¹⁰.

How this call fails

It is 2030 and the entry rung recovered. The likeliest cause is that the 2026 gap was mostly cyclical. The Canaries authors never claimed AI caused it². A second path runs through review work. Half-reliable agents created more checking than firms planned for¹. AI ROI leaders already added juniors in 2026³. The rest followed once seniors grew scarce and expensive¹¹. A third path is substitution. Apprenticeships absorbed most of the UK trainee decline¹². A new entry route could keep young workers in exposed jobs under a different title.

Who loses

• Graduates from 2026 to 2029 carry debt priced on a degree premium that fell from about 87 to 78 percent¹³.
• Aspiring trainees at elite firms lose the most places¹².
• Firms that plan to buy seniors externally will pay a rising experience premium¹¹³.

First moves

• Size junior cohorts to the review work that half-reliable agents create¹.
• Write junior training into fixed-fee bids as a priced line⁵.
• Grade credited learning hours on legal and financial judgment as well as tool fluency⁹.
• Graduates should favor employers that fund simulation and mentorship⁷.
• Press legislators to extend disclosure from layoffs to entry hiring¹⁰.

What would change our mind

• A Canaries revision with the young-worker gap below 10 percent would break the forecast².
• A flat METR horizon through 2027 alongside flat fixed-fee adoption would weaken both drivers¹⁵.
• Passage of the SKILL Act followed by a measurable rise in graduate hires would show policy can hold the rung open¹⁰.
• Most large firms keeping funded training budgets through a down year would refute our second claim⁷.