AI is already changing the workforce, but it’s not changing equally. Despite real momentum behind broad AI literacy efforts and the UK government’s push for apprenticeship routes into digital skills, a two-tier workforce is taking shape inside organisations, divided between those who’ve embraced AI and those who haven’t.
WRITER’s latest AI Adoption in the Enterprise report found that 92% of C-suite executives say they are actively cultivating a new class of AI “power-users” within their organisations. These super-users, roughly 4-5% of the workforce, are already reaping career rewards: they’re around three times more likely than their peers to have received a promotion or a pay raise in the past year.
AI fluency has become the floor, not the ceiling, and the consequences of operating below it – in productivity, career trajectory and workplace relevance – are compounding every month.
That split won’t close on its own. And throwing more tools at the problem won’t fix it either. The organisations closing this gap are treating AI as a people-transformation first, activating early adopters to build the scaffolding and use cases and creating the conditions for the whole workforce to follow.
The AI usage gap
These super-users tend to be concentrated in functions like marketing, HR, sales and customer support. And critically, they’re not just moving faster on the same tasks or drafting emails a little quicker. They’re building custom workflows from scratch, automating entire workflows and picking up skillsets that didn’t exist in most job descriptions two years ago, like systems thinking, workflow design and agent orchestration. That’s a meaningfully different kind of work, and it’s creating meaningfully different career outcomes.
Consequently, the pressure on everyone else is rising quickly. A total of 95% of executives say they’re already reorganising departments, teams and roles around AI, and 77% say that employees who don’t become proficient simply won’t be in the running for promotions, raises or leadership roles. Meanwhile, 60% go further, saying they plan to let go of workers who can’t or won’t adopt AI at work. What employees hear in all of that is a pretty clear message: figure it out on your own or get left behind.
Unsurprisingly, that message is backfiring. It’s producing resentment, disengagement and active resistance, with as many as 29% of employees admitting to working against their company’s technology strategy, with fear of job loss as the primary driver. Among Gen Z, the number jumps to 44%.
What leaders need to do differently
The gap isn’t accidental. It’s the result of how most organisations have approached AI from the beginning. Across the board, AI strategy hasn’t kept pace with AI investment. WRITER’s report found that 39% of executives admit they don’t have a formal strategy in place to drive revenue from AI, and 75% say whatever strategy exists is more performance than practice. Without clear direction from the top, adoption fragments by default, and the people with the most appetite naturally pull further ahead of everyone else.
Closing that gap requires a much more intentional approach than most organisations have attempted so far.
The organisations seeing real returns are treating their early adopters as multipliers rather than outliers. They’re getting these super-users together in a room and saying, “pretend the rest of the org chart doesn’t exist. Build the cleanest, most agentic process possible”. Then they’re empowering this group to document those use cases, share their workflows and give the rest of their organisation real examples of exceptional, autonomous work.
Critically, these leaders are ensuring that their power-users feel confident in reshaping their own roles for the age of agentic AI, not anxious that doing so makes them a target for replacement.
But spreading those gains requires sustaining momentum past the initial rollout. The organisations pulling ahead are building AI into how work gets evaluated, celebrated and improved on an ongoing basis, not just announcing a rollout and hoping habits form. That means creating space for employees to share what’s working, whether through internal showcases on a team all-hands or cross-functional use case libraries, and hosting monthly hackathons and workshops.
When people see what their colleagues are building with AI and see the concrete impact, adoption stops feeling like a mandate and starts feeling like an exciting opportunity.
The third is making governance visible. When employees don’t know what an agent can access, who approved it or who’s accountable if something wrong, the rational response is to avoid it or work around it entirely. Enterprises need clear policies on access, accountability and escalation paths in order to give AI any real authority over real workflows, with the confidence that when something goes wrong, you’ll know immediately, and you can fix it. Governance is not the enemy of speed. It’s the precondition for it.
Today’s two-tiered workforce
When organisations treat adoption as an individual responsibility and offer no real strategy to scale, you don’t get broad-based AI fluency. You get a small cohort of super-users and a much larger group that’s anxious, disengaged and quietly opting out.
Every month without a strategy is a month in which the gap grows harder to bridge. And as we all know, the best time to start was yesterday, but the second-best time is today.
Jevan Soo Lenox
Jevan Soo Lenox is chief people officer at WRITER. He previously held chief people officer roles at insitro and Stitch Fix, both companies using AI to transform human work and industry practices in drug discovery and e-commerce respectively. He has also held leadership roles at Square, Blue Bottle Coffee, and Minted.
Jevan began his career as a management consultant at McKinsey & Company at their Boston, New York, and Shanghai offices, serving clients across the technology, healthcare, and education sectors on strategy, human capital and organisational effectiveness as well as leading McKinsey’s Asia-Pacific recruiting strategy across 12 countries.
