The Two-Speed Workforce
A culture problem nobody anticipated
We’re seeing a new workplace culture challenge that didn’t exist a year ago.
Inside companies that have invested in AI tools and made them available across the workforce, a split is forming. A small group of employees has changed how they work entirely. They’ve moved past using AI to draft emails or summarize documents. They’re building automations, writing code, compressing entire project timelines, producing work that would have required a full team not long ago. For them, AI is how they operate. They talk to their computer and it executes complex tasks behind the scenes. They focus on decisions and outcomes instead of process.
Then there’s the majority of their coworkers who have the same tools, but for a variety of reasons, they’re using AI at a surface level or not at all.
Most organizations aren’t experiencing this conflict yet. Most companies haven’t equipped their workforce with the kind of AI tools that create this gap. But that’s going to change. We expect that within six to eight months, the tools now available to a small number of organizations will be widespread. And when they are, every company will face this issue.
That’s why we’re writing about it now.
The Conflict
When someone discovers they can produce a competitive analysis, a first draft, or a full project plan in an hour, their sense of what’s reasonable shifts permanently. Six months ago, a two-week turnaround on a deliverable would have felt normal. Now it feels like a bottleneck. Not because the deadline changed, but because their capabilities did.
The employee operating at AI speed starts to experience everything around them differently. Meetings feel slower, and handoffs take longer than they should. Projects that depend on contributions from colleagues who haven’t made the same shift start to stall in ways that feel avoidable. The frustration builds. And it changes how that person views the people around them, the pace of the organization, and whether this is still a place where they can do their best work.
On the other side, the colleague with the same tools on their desktop watches all of this happen and doesn’t know what to make of it. Some feel threatened. Others get defensive or quietly disengage. None of this is because they lack access. Something else is in the way: maybe it’s uncertainty about where to start, maybe a manager who hasn’t signaled that this is how work should be done now. Or maybe they are just modern-day Luddites, but instead of smashing weaving machinery, they just refuse to use AI tools.
The gap compounds quickly and research backs this up. Studies show that as AI-powered employees produce more, implicit pressure spreads across the team. Workers report that AI has increased their workload, not reduced it. Data from ADP Research shows that daily AI users report the highest engagement and motivation at work, but also weaker connections to their coworkers. They love the work, but they’re drifting from the people.
We asked John Trainor, President of Four Technologies, what this dynamic looks like from where he sits.
“I need to figure out how I can unlock the minds or activities of the people who have not yet figured out how their work can be better and faster with AI. On the flip side of that coin, I will also say that I need to know who is using AI and make sure that they are using it responsibly. I need to take that into consideration when I rely upon work that has been produced by someone who is leveraging AI, but maybe is not yet skilled in understanding what the inherent risks of AI might be.”
Leaders like John are managing in both directions at once, moving non-adopters forward while ensuring power users produce work and stay engaged. That’s the leadership challenge coming soon to every organization.
What Leaders Can Do
The instinct for most will be to treat this as a training problem. Run a workshop. Send a how-to guide. Encourage people to experiment. That helps, but it doesn’t address the cultural dimension of what’s happening.
The single biggest predictor of whether an employee crosses the threshold from surface-level AI use to real adoption isn’t access or aptitude. It’s their direct manager (as it has always been for almost every workplace challenge). When managers actively use AI themselves and make it clear that this is how work gets done now, adoption follows. When they don’t, it doesn’t. That’s the lever.
Here’s what we’re telling leaders who are starting to see this split form:
Call it out. The two-speed workforce is real, and pretending it will resolve on its own gives it time to harden. Acknowledge to your teams that adoption is uneven, that this creates tension, and that you’re paying attention.
Make AI part of how you lead, not something you’ve approved and moved on from. If your people see you using these tools to make decisions, prepare for meetings, and move work forward at incredibly new speeds, they’ll follow, we assure you. This is what we’ve done internally at NICH, where we’ve fully transitioned to an AI-first work methodology.
Rethink timelines and workflows. This is the practical one, and it’s harder than it sounds. If some of your people can now do what used to take weeks in hours, your project planning needs to reflect that without punishing the people who aren’t there yet. You’re not managing a performance problem. You’re redesigning how work moves through your organization.
Your AI power users are your most curious, growth-oriented employees. They’re also the most likely to leave if they feel held back by the pace around them. They know they are creating an entirely new skill set that will make them highly marketable to other power-user organizations. If you’re not paying attention to how they’re experiencing work right now, you could lose them before you realize there was a problem.
This is Culture Work
We’ve spent a decade at NICH helping organizations see the gap between how leaders think their culture is functioning and what employees actually experience. The two-speed workforce is a new version of something we’ve seen before: a shift happening on the ground that leadership hasn’t fully registered yet.
The difference this time is the speed. This gap will widen faster than anything we’ve encountered in our work. The organizations that get ahead of it will keep their best people and bring everyone else along. The ones that wait will be reacting to a problem that’s already set.
We’re building something that we think will help organizations see these dynamics clearly. More soon.
Links:
OpenAI: The State of Enterprise AI The report that quantified the gap between AI power users and everyone else, even inside the same companies with the same tools.
HBR: AI doesn’t reduce work, it intensifies it UC Berkeley research on what happens when AI-powered productivity creates pressure across the organization.
Gensler Global Workplace Survey 2026 AI power users spend less time alone and more time learning. Challenges the assumption that AI isolates workers.
WEF: AI is becoming your new work colleague ADP data showing daily AI users are more engaged but less connected to coworkers.
The two-speed split is real, and inside regulated companies, a layer of it is completely invisible. Some people are building AI fluency on personal tools after hours — prototyping workflows, building analytical frameworks, and developing system-building skills that their company doesn't know exist. Others are waiting for IT to roll out the next approved version.
The direct manager point is the one I keep coming back to. Inside my environment, the manager determines whether "I used AI for this" is a credibility signal or a credibility risk. That single variable determines whether the invisible practitioners stay invisible or start sharing what they've learned. The disclosure incentive is the bottleneck, and most AI adoption strategies don't even see it.