AI-driven layoffs are starting to look short-sighted and costly
Four lessons for businesses and government contractors
AI is constantly in the news.
However, the drumbeat about AI eliminating jobs and causing mass unemployment has given way to a sobering new reality: AI-driven layoffs may have gone too far—and for good reason.
According to Gartner, half of the companies that cut workers for AI-related reasons will hire those roles back by 2027. Forrester’s Predictions 2026 report had already documented the underlying cause: Fifty-five percent of employers who restructured for AI now regret the decision. Ford’s recent decision to rehire engineers who were laid off due to AI also has made headlines. Ford has spent the last three years hiring back more than 350 veteran engineers — “gray beards,” internally — to catch vehicle quality problems its automated systems missed. The automation issues cost the company billions.
All of this has prompted many to begin wondering about how do you know the value of AI before you implement changes, such as organizational restructuring.
Watch the Government Contractors webinar entitled, “How Can You Get Measurable Value from Artificial Intelligence (AI)“ .
The explanation from Charles Poon, Ford’s VP of vehicle hardware engineering, is the part worth sitting with: the engineers had left before their insights could be used to train the technology. “Artificial intelligence is a fantastic tool, but it’s only as good as the information you use to train it,” he said. “Mistakenly, we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that would produce a high-quality product.”
Ford is not an outlier. Gartner predicts that by 2027, half of all companies that attributed headcount reductions to AI will rehire staff for similar functions — often under different job titles. Forrester found that 55% of employers who restructured for AI already regret it. IBM’s AI handled roughly 94% of routine HR requests and choked on the other 6% — the ethical judgment calls — and IBM is now tripling U.S. entry-level hiring.
Here is what the coverage keeps getting wrong: this is not an AI failure story. Ford didn’t quit AI. The returning engineers rebuilt the AI tools and now lead troubleshooting, and CEO Jim Farley credits the effort with hundreds of millions in savings.
In short: Humans make AI work.
Where’s the failure? It’s a lack of measurement and not assessing where your technology and teams are, and not identify what value from AI looks like or what success looks like. Reducing payroll and adding value are two different measurements, and unfortunately, many companies only focus on the former (often to satisfy the old way of doing things on Wall Street) rather than the latter.
For government contractors, here are four lessons to consider:
1. Stop selling FTE reduction. Start selling measured outcomes.
If your proposal promises 30% labor savings from AI, and Gartner’s reversal rate holds, you have just written yourself a re-scope in year two and a past performance problem in year three. Headcount is an input metric. It tells a CO what you spent, not what you delivered. Bid the outcome — cycle time, error rate, protest exposure, award rate — and instrument it from day one.
2. Your gray beards are the differentiator, not the cost problem.
Ford paid billions to relearn this. The judgment your capture lead has about a customer, an evaluation panel, or a compliance regime is exactly the thing a competitor’s model cannot retrieve — because it was never written down anywhere a model could reach. That’s not overhead. That’s the asset. Price it accordingly, and capture it before you restructure around it, not after.
3. Protect the entry pipeline.
IBM’s CHRO Nickle LaMoreaux put it plainly: “If we don’t continue to invest in entry-level hires, what happens in three-five years? There’s no pipeline; the well simply dries up.” GovCon is more exposed than most industries here. Cleared, experienced staff are already the binding constraint, and you cannot import them. If AI absorbs the analyst-level work where judgment gets built, you have no seniors in 2031 — and no way to buy your way out.
4. Assess before you cut, because everybody else skipped it.
That 55% regret figure is a market. It is also a warning. The organizations now reversing course didn’t undervalue experience — they took a shortcut. They cut without ever mapping where judgment actually lived in the work. You can’t find that on an org chart. You find it in the workflow: walk one process end to end, and at each step ask what triggers the next action, who or what performs it, and what judgment it takes to move on.
Two weeks of that would have saved Ford three years.
We’re running a webinar on Wednesday, July 22nd entitled — How Can You Get Measurable Value from Artificial Intelligence — covering baselines, task-level scoring, and the readiness questions to answer before you automate anything. In the meantime, I recommend you check out our AI Services, led by Scott Andersen, which can be viewed at: https://www.governmentcontractors.com/ai-services/