The short version: AI is not cutting headcount through layoffs; it shows up as reduced hiring into entry-level roles, which is where the judgment work of the future is grown. Abacum lets you model the hiring plan and the AI assumption as separate drivers, so the trade-off is visible while you still have time to decide it.
What a named CEO put in an SEC filing
In August 2026, Etsy cut about 220 roles. The staff memo was filed as an exhibit to an 8-K, so the whole thing is public. On timing: "We’re also getting ready to plan for 2027. We didn’t want to ask teams to build those plans around an organization we already knew needed to change." On cause: "Second, these decisions weren’t driven by AI."
That is not an AI-skeptical memo. The same document says AI "is changing how all of us work, and it will continue to change how we build products and solve problems." The CEO simply declined to use AI as the justification for a headcount decision.
The displacement is real, but the mechanism is not what most plans assume
Challenger, Gray & Christmas recorded AI as the cited reason in 3,462 announced US job cuts in August 2026, the lowest monthly total since December 2025, ending a five-month run as the single most cited reason. Across 2026 so far, AI remains the leading reason given, named in 116,175 cuts, roughly 22 percent of the total. Both facts are true at once.
Stanford’s Digital Economy Lab, working from payroll records through June 2026, found no evidence of widespread, economy-wide job displacement. What it did find is that employment of 22- to 25-year-olds in AI-exposed occupations sits about 19 percent below where it would be if it tracked their less-exposed peers. The paper describes the mechanism directly: it "operates primarily through reduced hiring of young workers rather than increased separations." The authors call these early descriptive indicators rather than causal estimates.
Put those together, and the picture gets more specific. AI is not removing people you already employ. It is quietly removing the case for the next junior hire, one requisition at a time.
How does this inform your hiring plan?
Here is the test I would use. Take every open req and put it in one of two columns.
Drafting work: pulling data, writing commentary, building a deck, writing a blog post. An agent genuinely compresses that work, and the gain there is real.
Judgment work: approving the numbers, catching issues, defending it to the board, differentiating from competition, building a brand. This requires experience, expertise and domain knowledge that is learned over time.
The long term issue I see: teams are cutting from column one, because those are the junior seats and the easiest and least risky to automate. However, if every company is cutting from column one, who is investing in junior talent and growing them to be able to accomplish judgment work in the future?
It is genuinely interesting to see how this trend may play out and what it means for hiring. If AI absorbs entry level jobs, where do new people enter the workforce, and how does training and enablement need to change?
As you create this year’s headcount plan, challenge your team to think further into the future, and about who you may need to acquire to make sure you do not have talent gaps in the long term. Our free headcount planning template separates filled, offered and open seats so the entry-level line is visible instead of buried in a total.
How Abacum helps
Software has genuinely taken work out of the first draft. It has not removed the person who signs their name under the number. Abacum gives you somewhere to hold the assumption explicitly. Model the hiring plan and the AI assumption as separate drivers, see what each one does to revenue, cash and margin, and keep the roster current from the systems your people data already lives in. The headcount planning guide for finance teams covers the wider cycle.
