AI Job Cuts Aren't Improving Return
Most companies that cut jobs to make way for AI saw no financial payoff. We are measuring the wrong things, again.
Recent data from Gartner has exposed a hard truth in the tech industry: 80% of companies that cut jobs to make way for AI initiatives didn’t see any correlation with improved financial returns. In fact, organizations that make deep, premature cuts risk losing institutional knowledge and face a high likelihood of having to make costly rehires down the road. So, what’s really going on?
If you ask me, we are dealing with a perfect storm of two distinct factors. First, we are operating in a market heavily clouded by economic uncertainties. Second, the sheer, breakneck speed of AI’s growth has left leadership teams deeply confused about how to actually deploy it effectively.
The Metrics Trap: Repeating Dot-Com Mistakes
Because companies are confused about integration, they are retreating to vanity metrics. Right now, some managers are measuring productivity by tokens generated, while others are tracking raw lines of code.
Does this sound familiar? It’s exactly what we did during the dot-com era before the bubble burst. Back then, we measured “eyeballs” and page views instead of actual revenue and retention. Today, we’re mistaking code volume for value. We need to stop measuring developers by how many lines they churn out and start measuring them on quantifiable output and, better yet, clearly articulated business goals.
AI is a Tool, Not a Business Engine
Here is the reality: AI is an incredible tool, but on its own, it cannot grow a business. A language model can generate a script, but it doesn’t understand your customer’s underlying pain points, nor can it strategize a product rollout.
AI’s true power lies in automating the tedious, repetitive parts of the job. It should be used to eliminate drudgery, not to eliminate a developer’s livelihood. When we use AI to do the same work with fewer people instead of unlocking new capabilities, we optimize for the smallest possible version of what the technology can deliver.
Protecting the Craft of Development
It is absolutely vital for developers to stay in the trenches. They need to debug, architect systems, and get hands-on experience writing code. That struggle is where institutional knowledge and deep problem-solving skills are forged.
AI should be available right alongside them, a powerful assistant ready to help when they get stuck, need to identify an elusive bug, or want to refactor a messy function. We must build operating models where human developers guide autonomous systems, rather than allowing the craft to be hollowed out entirely.
Let’s stop managing for the optics of efficiency and start managing for actual growth.