I just read the latest Harvard Business Review study on AI-driven layoffs and I can’t stop thinking about one finding:
“Companies are laying off workers based on AI’s potential, not its proven performance in their specific context.”
Then I saw another report: 55% of employers regret their AI-driven layoffs.
Let me connect these dots from a product and strategy perspective, because I think we’re witnessing one of the most irrational business decisions of the decade.
The Optics vs. Reality Problem
I work in product strategy. I attend board meetings. I hear the conversations that lead to layoffs. Here’s what I’m seeing:
What boards hear:
- “OpenAI just released GPT-5, productivity up 10x”
- “Our competitor announced they’re ‘AI-first’”
- “McKinsey predicts AI will replace 30% of knowledge work”
- “If we don’t act now, we’ll fall behind”
What’s actually true in our organization:
- AI tools improve productivity by maybe 15-20% for specific tasks
- Our “AI features” are mostly wrappers around OpenAI APIs
- Customers aren’t asking for AI—they’re asking for reliability
- Engineers spend more time debugging AI outputs than they save
But the board narrative is: “We’re transforming to AI-first, reducing headcount by 20%.”
The actual result: We laid off experienced people and hired fewer, more expensive AI specialists who are now overwhelmed.
The Herd Mentality Hypothesis
I’ve been tracking tech layoffs. Look at the pattern:
Q4 2025:
- Meta announces “AI efficiency gains” layoffs
- Stock goes up 8%
Q1 2026:
- Google announces similar “AI transformation” layoffs
- Stock goes up 6%
Q2 2026:
- Every Series B+ startup announces “strategic AI realignment” layoffs
- Investors applaud “operational discipline”
Is this rational business strategy or are we just copying what looks good to investors?
I suspect it’s the latter. Companies aren’t laying off because AI has proven it can replace those workers. They’re laying off because:
- It makes good PR: “We’re innovative and AI-forward”
- Investors like it: Signals you’re cutting costs and embracing future tech
- Everyone else is doing it: FOMO that if you don’t, you’ll look behind
But none of this is based on actual AI performance in their specific business context.
The Maturity Gap: Promises vs. Reality
Here’s what AI tool vendors promise:
- “80% reduction in coding time”
- “Automate customer support”
- “Replace junior analysts”
- “10x productivity gains”
Here’s what we’re actually seeing in our organization:
AI Coding Assistants:
- Promise: 80% faster development
- Reality: 15-25% faster for boilerplate, slower for complex logic, creates bugs that take time to fix
- Net gain: Maybe 10-15% productivity, highly variable by task
AI Customer Support:
- Promise: Automate 70% of support tickets
- Reality: Handles 30% of simple tickets, escalates 70%, customers frustrated by bot interactions
- Net result: We still need most of our support team, plus engineers to maintain the AI
AI Data Analysis:
- Promise: Replace junior analysts
- Reality: Generates insights that need senior analyst validation, misses context, sometimes hallucinates data
- Net result: Analysts spend time validating AI instead of doing original analysis
The gap between promise and reality is massive. Yet companies are laying off based on the promise, not the reality.
The 55% Regret Number
Let’s talk about that HBR finding: 55% of employers regret AI-driven layoffs.
Why are they regretting it?
From conversations with peers:
-
AI didn’t deliver promised productivity: “We thought AI would replace 5 people. It replaced 0.5 people’s work.”
-
Lost institutional knowledge: “We laid off someone who’d been here 8 years. Turns out they were the only person who understood our legacy billing system.”
-
Remaining employees burned out: “The people we kept are doing the work of 3 people. They’re quitting.”
-
Quality degraded: “AI makes mistakes. We used to have people who caught those mistakes. Now we don’t.”
-
Customers noticed: “Our support response time doubled. Our bug rate tripled. Customers are churning.”
But the damage is done. You can’t easily rehire the people you laid off. They’ve moved on, lost trust, or left the industry.
Is This Rational Business Behavior?
Let me apply a basic strategy framework:
Rational layoff decision-making:
- Identify specific work that’s redundant or low-value
- Verify that AI can actually do that work (proof of concept)
- Test AI solution for 3-6 months
- Measure actual productivity impact
- Make staffing decision based on data
What’s actually happening:
- Read news about AI capabilities
- Assume it applies to your business
- Announce layoffs to please investors
- Figure out AI implementation later
- Discover AI doesn’t work as promised
- Regret decision but can’t reverse it
This isn’t strategy. This is hype-driven decision making.
The Human Cost
I want to be clear about what this looks like on the ground:
Real people I know who were laid off “for AI”:
-
Sarah: 12 years in customer support, laid off because “AI will handle tier-1 support.” The AI handles 20% of tickets. The team is overwhelmed.
-
James: Senior data analyst, laid off because “AI will automate reporting.” The AI generates reports full of errors. Nobody validates them anymore.
-
Maya: Technical writer, laid off because “AI can write documentation.” The docs are now AI-generated and full of inaccuracies. Engineers complain they’re unusable.
These aren’t “low performers” who needed to go anyway. These are competent professionals who were sacrificed to an AI narrative that hasn’t materialized.
The psychological damage:
- Everyone remaining is terrified they’re next
- Productivity drops from anxiety and overwork
- People start looking for jobs at companies that aren’t “AI-first”
- The best talent leaves proactively
The 12-18 Month Correction Hypothesis
Here’s my prediction: In 12-18 months, we’ll see a wave of “quiet rehiring” as companies realize AI didn’t deliver.
It won’t be announced as “we were wrong.” It’ll be framed as:
- “We’re scaling our AI teams and need support staff”
- “We’re investing in customer experience”
- “We’re building our next-generation products”
But what’s really happening is companies realizing they over-corrected.
The problem: The people they laid off won’t come back. They’ve found other jobs, changed careers, or lost faith in tech. The industry will pay a premium for new people who won’t be as experienced.
What Product Leaders Should Do
If you’re in product or strategy and facing pressure to justify AI-driven layoffs:
1. Demand proof of concept before headcount decisions
- “Show me AI working in production for 3 months before we lay people off”
- “Prove the productivity gains are real, not theoretical”
2. Separate cost-cutting from AI narrative
- If layoffs are for financial reasons, say so
- Don’t blame AI for business decisions
3. Measure actual impact, not promised impact
- Track productivity before and after AI implementation
- Compare AI outputs to human outputs on quality, not just speed
4. Consider augmentation vs. replacement
- Can AI make existing employees more productive?
- Is “humans + AI” better than “AI replacing humans”?
5. Factor in hidden costs
- Time spent validating AI outputs
- Lost institutional knowledge
- Decreased employee morale
- Customer satisfaction impact
The Question I Can’t Answer
Is it ethical to lay off workers for technology that doesn’t yet work as promised?
If a company says “We’re laying you off because we have a new technology that can do your job,” but that technology is unproven in their context, is that honest?
Or is it just a convenient narrative to justify cost-cutting that leadership wanted to do anyway?
I don’t have the answer. But I think we need to have this conversation more honestly.
For others in product, strategy, or leadership: Are you seeing the same pattern? Are the AI productivity gains real or are we in a hype cycle that will crash when reality sets in?
I’m genuinely curious whether I’m being too cynical or whether this is a emperor’s-new-clothes moment that we’ll look back on in embarrassment.