The Numbers That Matter
Anthropic just published data on what AI agents did to its own workforce. Junior technical hires dropped from 27% to 9% of new technical hiring since January 2026. The share of junior employees in technical departments fell from 21% to 14% over the same period.
The shift started in teams that used Claude most: code writing and bug hunting moved to agents first. By April 2026, a majority of managers said they were hiring for different skills because of AI. About a third of ML managers and a fifth of software engineering managers said they were hiring less because of AI.
Anthropic states directly that its redeployment programs help current employees but do nothing for the junior engineer who never got an offer.
What This Means for Sales Teams
The report matters for sales because Anthropic tripled headcount anyway, but the mix changed. Technical departments fell from 61% to 56% of total headcount. Sales, finance and recruiting grew faster than engineering in absolute numbers.
Kelly Loftus scaled Anthropic's sales org from under 10 people to 150+ in 18 months while total headcount went from 250 to 1,300. Her team runs without traditional quotas, using shadow targets instead. Eleanor Dorfman rebuilt the sales org in January 2026, and four months later 54% of new enterprise logos were coming in self-serve.
That self-serve number is an outlier driven by extreme demand. Do not plan your pipeline against it.
Agent sessions at Anthropic now work six times more hours than humans, but spend 80% of their time waiting on human review. Code commits per engineer run at 5x their level from six months earlier. An executive assistant who built her own scheduling agent said entry-level admin work was already automated, and her job had moved to chief-of-staff work.
The ANZ Angle
Anthropic competes in ANZ through global enterprise and cloud channels rather than a dedicated local org. Claude sits at roughly 40% of enterprise LLM spending globally, ahead of OpenAI at 27%, according to Menlo Ventures data. The workforce shift suggests customers should expect similar headcount planning conversations: fewer junior hires, different skill requirements, heavier agent deployment.
The company's annualised revenue passed $47 billion in May 2026 at a $965 billion valuation. Revenue per employee runs 2-4x what Google and Meta did at year five, adjusted to 2025 dollars. Administrative roles barely grew while the company tripled, driven by agent assistants and automated review.
For sales leaders: the bottleneck is human review time, not agent output. Reading everything agents produce would take technical teams more than half the work week. That constraint shapes what gets automated next.