Loading today’s stories
Leadership1 distinct publisher3 min readPublished
If the projections hold, most employers spend this decade short of people while being asked for roughly a fifth more output, and a workforce plan that budgets headcount down through automation answers the wrong constraint.
The Board Room · Leadership desk
Compiled by The Board RoomSomething wrong?How this is made
The two projected rates only make sense read against each other. If output rises 22% while the number of workers rises 3.5%, output per worker has to rise about 18% over the decade, since 1.22 divided by 1.035 is 1.179 [20], which compounds to roughly 1.7% a year [21]. Josh Bersin, whose reading of the release this piece works from, puts the required figure at almost 2% annually and treats it as grounds for accelerating AI programs [3]. Both numbers describe the same shape: demand for output grows about six times faster than the supply of people to produce it [22]. An AI budget defended purely as cost takeout is being argued in the wrong category, because under these projections it is a capacity purchase.
The composition matters more than the totals. Bersin's grouping of the BLS taxonomy's 830 job titles into four categories puts around 61% of jobs in physical or human-centered work [19][4], and estimates that AI can today fully automate roughly 13% [5]. That 13% is largely the administrative layer, the cashiers, clerks, bookkeepers, secretaries and tellers whose titles are shrinking fastest [15], and it now reaches some IT roles including support specialists, testers, non-developer programmers and data entry [16]. Against that, healthcare supplies 37% of projected job growth [14], and nurse practitioner pay is projected to move from $132,000 to $181,000, a rise of $49,000 or 37% [13][23]. Savings scraped from a low-wage clerical base do not buy labor in a market where the scarce roles are repricing upward.
A skeptic would say a ten-year projection is a trend line, and that the automation assumption buried inside it is the least reliable part of the model. That is fair, and it cuts a particular way. The supply side of this projection rests on aging, birth rates and immigration policy [6], which move slowly and are visible years in advance, and The Economist's view, as cited by Bersin, is that current US immigration policy makes the shortage worse [7]. The automation side is where the record is genuinely thin: 13% is an estimate of potential rather than adoption [5], and the four-way sort of job titles is Bersin's construction, not the BLS's [19]. We do not know the adoption path.
This quarter's version of the decision is whether to write automation savings into next year's headcount plan. The consequence arrives in the market for older workers, where the over-65 share is projected to go from 7% to 8.5% of the workforce [8], which on a workforce 3.5% larger works out to about 26% more people over 65 [26]; AARP finds almost 10% of over-65 workers unretiring [9], while nearly a quarter of workers over 50 believe they have been left behind [10]. Bersin calls the increase seven times the overall growth rate, though 21 divided by 3.5 is six [25].
The asymmetry settles it. Money spent on retention and skills is recoverable if automation lands harder than the BLS expects, because the same people operate the tooling; the management layer that such plans usually thin first is a layer the BLS still has growing [17]. A headcount plan built on automation that arrives late ends with the firm buying scarce labor at 2030 prices.
Ranked by verification strength, evidence, and original report placement.
Josh Bersin, summarising the new Bureau of Labor Statistics Employment Projections 2025-2035, writes that the US labor market will grow by only 3.5% in ten years while GDP is expected to grow by 22%.
The projected 3.5% workforce growth compares with 10.9% growth in the last decade, which Bersin describes as almost a third slower.
Nurse practitioner jobs are projected to grow 41% in number, with average wage going from $132,000 to $181,000.
Bersin gives three reasons for the labor slowdown: anti-immigration policies, low birth rate, and aging.
Bersin writes that the US administration's anti-immigration policy is further constraining the workforce and that The Economist notes this will exacerbate the workforce shortage.
Studies by AARP show almost 10% of workers over 65 are unretiring, a trend Bersin expects to continue and to lengthen careers.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · September 2, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
leadership
Retention plans keep pricing a loyalty that has been draining for three decades1 distinct publisher
product
Rillet's $100M reads as proof mid-market ERP is rip-and-replace, mostly at the cheap end1 distinct publisher
leadership
Disney swaps raises for discounted stock and a full health-plan re-enrollment1 distinct publisher
leadership
AI's answer keys are being written for $85 an hour by people who cannot get other work1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Public dataset, private arithmetic
Two different kinds of number are mixed here without a seam. The 3.5%, the 22%, the 41% growth in nurse practitioner roles and the 13.3%-to-14% healthcare share come from a federal projection release that exists and can be audited — though our coverage paraphrases rather than shows it. The shares that carry the argument, 61% of jobs as human-centred and 13% as fully automatable, come from Bersin's team sorting 830 job titles into four buckets with no method, threshold or definition published. And the internal figures do not reconcile: the productivity rate is rounded up and the over-65 multiple is overstated.
Nothing shipped to measure
This is a ten-year projection and an argument about it. No employer redesigning roles on the record, no spending or hiring disclosure, no tool in production, no dated event of any kind. Adoption here is not weak — it is simply outside what this reporting contains, and inventing a reading would be worse than leaving it open.
Reassurance asserted, displacement counted
'Despite the talk about job losses from AI, we now know the opposite is true' is a lot of certainty to rest on a projection — particularly one that, a few lines later, puts 13% of jobs within reach of full automation today and another 13% of employment in administrative titles that BLS expects to shrink faster. The displacement side arrives with percentages; the comfort side — reinvented work, nobody unemployed, Superworkers — arrives as assertion, alongside a book launch. The rounding drifts the same direction the argument does.
The author sells the answer
The reading that makes this decade a shortage rather than a cull is also the reading that sells advisory work, a named research construct, a longer report behind a link, and a book launching in October — and the piece says so, right down to telling HR professionals their market has never been hotter. That does not make the federal numbers wrong. It does mean the interpretation layered on top is being produced by the party who profits from that interpretation, with nobody else in our coverage checking it.
One reader of a public release
The direction of travel is credible and independently checkable in principle: federal projections do exist and a labour-constrained decade is a mainstream reading. Our confidence is capped by having a single voice, an interpretive layer nobody has reviewed, no primary tables in hand, and two arithmetic slips in the numbers we could test — which is enough to trust the shape of the story and not its decimals.