Everyone says AI adoption is an upskilling challenge. Train your people, and they'll adapt.
But leaders already use AI at four times the rate of individual contributors. Seventy-one percent of CEOs rank AI as their top investment priority. And yet ninety-five percent of organizations see zero measurable return on forty billion dollars of enterprise spending. Meanwhile, manager engagement has dropped nine percentage points since 2022.
What if the problem isn't that people can't learn AI — but that the way we're deploying it is building a structural divide?
Here's what we're noticing: organizations are facing a structural design problem. They're deploying AI without a coherent structural plan. And when you introduce a transformative capability into an organization without redesigning how work flows, the result has a name. It's the two-speed workforce.
What if the problem isn't that people can't learn AI — but that the way we're deploying it is building a structural divide?
The Access Gap
Who gets to use AI in your organization is not random. It follows the existing structure.
Leaders are the most likely to use AI to any extent — 69 percent, compared to 55 percent of managers and 40 percent of individual contributors. But the frequency gap is where the pattern becomes visible. Forty-four percent of leaders use AI at least a few times a week. For individual contributors, that number is 11 percent.[1]
This is a widening trend. Leaders' frequent AI use grew by about 25 percentage points between 2023 and 2025. Managers and employees saw much smaller increases over the same period.[2]
The industry split tells the same story. In technology and information systems roles, 76 percent of employees use AI at least occasionally. In retail, it's 33 percent. In healthcare, 37 percent. In manufacturing, 38 percent.[2] AI lands where the existing structure already concentrates decision-making power and technical infrastructure. It doesn't distribute itself evenly — it follows the path of least resistance.
AI follows existing organizational structures. And those structures concentrate access rather than distribute it.
Ninety-one percent of organizations say they use AI tools. Only 21 percent of workers actually do.[3] That gap — between what the organization claims and what the workforce experiences — is the access gap.
People on the front lines can learn AI. AI follows existing organizational structures. And those structures concentrate access rather than distribute it.
The Support Gap
Access is only the first pattern. The second is support — and here, the numbers reveal something about how organizations approach change.
Seventy-seven percent of employers say they plan to reskill workers for AI. Thirteen percent of employees have actually received any AI training.[3]
That's a structural choice — even if it's an unintentional one.
When employees do receive strong leadership support for AI adoption, the share who feel positive about generative AI jumps from 15 percent to 55 percent.[4] The difference is structure. Organizations that build support into the deployment see dramatically different outcomes. Those that don't are left wondering why adoption stalls.
And the training that does exist often misses the mark. Over half of professionals say AI trainings feel excessive — more like a second job than a capability builder.[5] This paradox — too little training, yet too much perceived burden — reflects poor pacing and design, not poor people.
When the structure doesn't support, fear fills the vacuum.
The consequence shows up in trust. Seventy-two percent of managers believe employees fear AI tools will make them less valuable at work. Seventy percent believe employees fear AI will lead to them being fired.[6] When the structure doesn't support, fear fills the vacuum.
The Outcome Gap
Here's where the pattern becomes most visible.
Despite roughly $40 billion in enterprise AI investment, 95 percent of organizations have seen zero measurable impact on profits. An academic survey of nearly 6,000 global executives reports that 89 percent see no effect on labor productivity.[7]
Why? Because AI initiatives in most organizations are concentrated at the task level. They arrive as individual tools, without connecting to a broader organizational redesign. Only 12 percent of employees in AI-implemented organizations say AI has transformed how work gets done.[8]
As Gallup puts it: "The gap between reported individual and firm-level productivity suggests that while AI is helping many employees work more efficiently, many organizations have not yet fundamentally redesigned workflows, roles or processes around AI."
Individual productivity gains don't compound into organizational transformation when they're concentrated in isolated pockets. The individual gains are real. The people using AI see measurable productivity improvements. But the people who don't — the half of the workforce that rarely or never uses AI — aren't just standing still. They're working within structures that haven't changed, alongside colleagues whose capabilities are accelerating in a different direction.
And the existing structures don't just persist — they get reinforced. The communication patterns that were already strained now carry an additional load: translating between those who have AI capability and those who don't. The decision-making processes that existed before AI now operate at two different speeds.
Individual productivity gains don't compound into organizational transformation when they're concentrated in isolated pockets.
The structure is beginning to show strain. Global manager engagement has dropped nine percentage points since 2022 — from 31 percent to 22 percent. The largest single-year fall occurred between 2024 and 2025.[9] Twenty-seven percent of employees in AI-adopting organizations say their workplace has changed in disruptive ways. In organizations that haven't adopted AI, that number is 17 percent.[8]
The two-speed workforce is an adoption pattern — and what happens when you introduce a transformative capability into an organization without redesigning the structure that holds it together.
What This Means
The two-speed workforce is what happens when organizations treat AI as a tool to be added, rather than a capability that requires structural alignment.
Pilots prove that AI works. But pilots don't prove that AI transforms. Transformation requires something pilots don't have: organizational coherence. A shared direction. A redesign of how work flows, how decisions get made, how capability distributes across the whole system — not just the parts that were selected for the experiment.
The question is whether the structure you're building can hold the capability you're introducing.
Here's something we're noticing: the organizations that will thrive in the next phase of AI adoption aren't the ones with the best pilots. They're the ones that treat AI as a structural decision — one that requires alignment on organizational goals, systematic change planning, and a design that includes everyone, not just the people who were chosen to go first.
The question is whether the structure you're building can hold the capability you're introducing.
Does that match what you're seeing?
Sources
- Gallup. “Rising AI adoption spurs workforce changes,” 2026 — Primary source for adoption gap data, leader vs. employee usage patterns
- HR Executive. “The AI adoption gap is real, and widening,” 2026 — Industry-level AI adoption variation, trend analysis
- The Network Installers. “AI in the workplace statistics & trends in 2026,” 2026 — Training gap data (77% plan vs. 13% received), organization vs. worker usage gap
- SUCCESS Magazine. “The AI adoption gap splitting leaders from teams,” 2025 — Leadership support impact on AI sentiment (15% → 55% positive)
- AI Work. “100+ AI in the workplace statistics you need to know in 2026,” 2025 — Training paradox data (over half say trainings feel excessive)
- Beautiful.ai. “AI’s impact on the workplace in 2026: 3rd annual survey of American managers,” 2026 — Manager perceptions of employee fear (72% fear less valuable, 70% fear fired)
- NBER Working Paper No. 34836 (Yotzov et al.), 2026 — Academic survey: 95% zero ROI, $40B investment, 89% no productivity effect
- Gallup. “Rising AI adoption spurs workforce changes,” 2026 — Transformation claim (12% say AI transformed work), disruptive change comparison (27% vs. 17%)
- Gallup. “State of the Global Workplace 2026” — Manager engagement drop (31% → 22% since 2022), largest single-year fall 2024-2025
Please note: 51&even is an AI-first organization. We embrace AI at every step of our value creation and build our processes with a deep integration of human-AI capability. Humans always have the last decision. But this text was heavily built with AI.