The AI skills gap is the widening disconnect between the technology enterprises can now deploy and the human capability required to operationalize it.
When every company has access to the same AI, what actually separates the ones that win?
There's a narrative floating around that AI is about to dramatically reduce the need for people. The Citrini Research 2028 Global Intelligence Crisis report puts a name to the fear: a world where AI-driven efficiency surges while the human labour economy lags behind.
After sitting with senior technology and operations leaders at our recent Client Advisory Board, I'd argue the opposite is happening. Nobody is struggling to find technology anymore. They're struggling to find the right people to make sense of it.
One comment stuck with me:"I have more tools than ever to make my job easier but I've never been busier."
Everyone laughed. But it landed because it was true for almost everyone at the table.
Automation platforms are mature. AI capabilities are accelerating weekly. Yet workloads aren't shrinking and expectations are expanding faster than teams can adapt. Bain's latest generative AI research found that 75% of companies are struggling to find the in-house expertise they need as they scale. What's emerging isn't a technology bottleneck. It's an AI skills gap.
One of the more provocative moments came when an executive said they currently value interns over senior developers. Not because experience isn't useful. Because adaptability suddenly matters more.
But the nuance the room landed on: that's not a reason to sideline experienced talent. Senior operators play a critical function right now, pressure-testing outputs and validating that what the technology produces is actually sound. The magic is in the convergence: unburdened curiosity paired with the guardrails of experience.
Another leader described hiring shifting from deeper to broader resources.
"The expectation is moving to a broader set of skills rather than a deeper set of skills. And that is harder to get."
As AI handles more technical heavy lifting, organisations need people who can move across domains and connect technology to business outcomes. The ideal profile right now is a translator with data fluency.
The BA function is becoming more valuable, not less. One participant offered a reframe worth borrowing: Business Engineer.
The scribing and documentation that ate half a BA's day? AI handles that now. What's left is what technology genuinely cannot replicate: reading the room, getting to root cause, and distilling lived experience into actionable requirements. Without that, automation overwrites complexity rather than accommodating it.
Even in organisations with clear process ownership, those owners can still have blind spots across end-to-end processes. Which led to one of the most honest observations of the day:
"I don't need more system data. I need the tribal knowledge."
AI struggles with the unwritten context experienced operators carry in their heads: why exceptions happen, why rules exist, what people actually do versus what the process diagram says. And people who feel uncertain about their future won't volunteer that context freely. Creating conditions where tribal knowledge gets shared, not hoarded, is a leadership challenge as much as a technology one.
AI isn't reducing effort yet. It's raising ambition. Teams aren't doing the same work faster. They're attempting problems they wouldn't have touched before.
The companies moving fastest aren't those with the newest tools. They're the ones closing the gap between what technology can do and how their organisation is set up to use it. Closing that gap requires rethinking your operating model: how you prioritise automation, how your architecture supports scale, and how you build human capability to move at pace. That's where the competitive edge lives.
