When the Companies Building the Tech Start Cutting Their Own Builders
Sign Up For Daily Job Market News
Visa confirmed today that it’s cutting 7% of its workforce, about 2,600 jobs, and the part that should get your attention is where the cuts land: technology and product teams. A company spokesperson confirmed the number, and CEO Ryan McInerney laid it out in a staff memo, framing it as the company continuing to “focus on driving efficiency” so it can reinvest in its highest-potential bets. Bloomberg broke the story and reported that AI has helped Visa cut repetitive tasks and speed up product development. A person familiar with the reasoning added the caveat that AI was “not the sole factor” behind the decision.
Watch that language, because it’s becoming the standard corporate script. AI gets named as a factor, then immediately walked back to “not the sole factor.” Companies want the efficiency story for investors without owning the “we replaced people with software” headline. The honest read usually sits in the middle: AI gives cover to workforce decisions that also involve overhiring, cost pressure, and a stock that’s been trailing the market.
Here’s the detail I keep coming back to: for years, the safe assumption was that technical roles were insulated and the automation risk sat with routine or clerical work. Visa just trimmed the exact teams that build its technology. And it’s not alone. Mastercard announced plans earlier this year to cut 4% of its global workforce, and Block said in February it would cut roughly 4,000 jobs, close to half its headcount. When the companies closest to the tooling start cutting their own builders, that’s a signal worth sitting with.
One more piece of context that changes how I read this. Visa had roughly 34,100 employees in 2025, a headcount that had grown 8% year over year. A 7% cut on the back of 8% growth reads less like a collapse and more like a correction. A lot of what’s being called an AI-driven layoff this year is really a company unwinding pandemic-era and post-pandemic overhiring, with AI as the convenient explanation. Both things can be true at once, and employers watching these headlines should separate the two before they take the wrong lesson into their own planning.
For hiring managers, the takeaway is that “safe from automation” is no longer a category you can assume. The roles that hold up are the ones where demand outruns supply, and knowing which of your roles those are is the whole game right now. If you’re weighing how to build a resilient team without over-rotating on efficiency, our take on how to hire starts from what the work actually requires, not from what a headline says is expendable.
Pro-Worker AI Isn’t a Fantasy… It’s a Funding Choice
While Visa was announcing cuts, MIT economist and 2024 Nobel laureate Daron Acemoglu was making a different argument in The Atlantic: the damage isn’t inevitable, and the direction AI takes is a choice we’re making right now.
His case runs like this: over the last 40 years, digital technology drove a wave of automation that raised productivity but funneled the gains to a small group while leaving a lot of workers worse off. Generative AI, he argues, is set up to repeat that pattern, only bigger, unless we deliberately steer it somewhere else.
That somewhere else is what he calls “pro-worker AI,” models built to make people better at their jobs rather than replace them. His electrician example is the one worth sitting with. You don’t need a chatbot that writes sonnets to troubleshoot electrical equipment; you need a specialized model trained on niche, high-quality technical data that can pull site-specific inputs like sensor readings and photos to diagnose a problem faster. Same idea in teaching, where a model trained on curriculum could spot patterns in test results, and in healthcare, where targeted tools could let lower-skilled workers safely take on more.
His worry is where the money goes. He points at the tax code: in the US, paying a worker $100 can carry up to $30 in tax and spending obligations, while $100 spent on automation equipment carries a tax burden under $5. That gap, he argues, pushes companies to automate even when a human would do the job better.
I like this framing because it moves the AI-and-jobs conversation off the tired binary of “robots take everything” versus “nothing changes.” The real question is what these tools get pointed at, and right now they’re mostly pointed at replacement, because that’s where the return is.
The electrician example lands hardest in staffing. The roles hardest to fill are skilled trades, healthcare, and technical field work, the jobs where demand has always outrun supply, and those are not the ones AI is racing to automate. A tool that makes a good electrician faster and helps a newer one ramp up sooner doesn’t threaten that worker. It widens the pipeline, which is exactly what those markets need.
Where I’d push back on the optimism: the tax point cuts both ways. Acemoglu treats that $30-versus-$5 gap as a distortion to fix, and maybe it is, but it also honestly describes why employers reach for automation first. That incentive is real, and it isn’t going away because an economist wrote a compelling essay. The companies that win the next decade will be the ones that make their people more valuable with these tools, not just cheaper to replace. That’s a management decision more than a technology one, and most of the deciding is happening now. If you’re trying to figure out which roles to grow and which to augment, that’s the conversation we have with clients through our staffing services.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
The Metric That Warns You Before Your Attrition Rate Does
If the first two stories are about what companies are cutting, the third is about how to know whether the people you keep are actually productive, and Gallup argues most companies measure it wrong.
The usual metrics (hours logged, units produced, revenue per employee) all measure output after the fact. By the time those numbers flag a problem, the conditions that caused it have been building for months. Gallup’s pitch is to measure the workplace conditions that predict productivity before it hits the financials, and its tool for that is employee engagement.
The data behind it is substantial. Gallup’s meta-analysis spans more than 183,000 business units across 53 industries and 90 countries. The headline: teams in the top quartile of engagement are 14% to 18% more productive than teams in the bottom quartile. The same gap shows up across other outcomes: 23% higher profitability, 78% lower absenteeism, 64% fewer safety incidents, and 41% fewer quality defects, and it holds up across industries and even through recessions.
The current picture isn’t good. In Q2 2025, only 47% of US employees strongly agreed they know what’s expected of them at work, down from 56% in 2020. Only 31% strongly agreed that someone at work encourages their development. The US engagement ratio fell to 1.8 engaged employees for every actively disengaged one in early 2026, while Gallup’s top-performing organizations run at 14-to-1.
The leading-indicator argument is the right one, and it’s the part hiring managers should take to heart. Revenue and turnover tell you what already happened. If you’re waiting for your attrition rate to spike before you act, you’re reading the smoke after the house is gone.
The stat that should stop people is role clarity dropping from 56% to 47% in five years. Fewer than half of American workers can say they know what’s expected of them at work. That’s a management problem more than a wellness one, and it’s the cheapest on the list to fix. Clear expectations don’t cost a dollar. They cost a manager’s attention.
