IBM's 2026 CEO study found that 86% of executives believe their employees already have the skills to work with artificial intelligence (AI). The same study found that only 25% of the workforce is actually using it regularly. I think that 61-point gap is the most expensive number in enterprise AI strategy right now, because it means most rollouts are stalling for reasons that have nothing to do with the technology itself.

Gallup's own workplace data point to why. Employees who use AI daily or several times a week are more than twice as likely to fear their job will be eliminated within five years, compared to employees who barely touch it. More usage does not automatically buy more comfort. Trust, clarity, and skills are what convert access into adoption, and leadership is the only thing that can supply the first two.

How Many Employees Actually Use AI at Work in 2026?

Gallup surveyed US employed adults between February 4 and 19, 2026, and found that half of them had used AI in their role at least a few times that quarter, up from 46% the prior quarter and 40% a year earlier. That's real, sustained growth. But total usage hides a much smaller number underneath it: only 13% of employees use AI daily, up from just 4% in the second quarter of 2024, and 28% use it a few times a week or more.

Layer IBM's finding on top of that and the picture sharpens. Executives at 86% of surveyed organizations believe their people have the skills to collaborate with AI. Only a quarter of the workforce is actually doing it on a regular basis. Licenses purchased and workflows automated are the numbers that make it into a board deck. They describe what leadership bought, not what employees actually do with it.

Every row on the left is an input. Every row on the right is the actual experience driving whether that input turns into use. Closing that gap is a leadership job, and it starts with asking a different question than most rollouts start with.

Is AI Adoption Really a Culture Problem?

Most leaders respond to disappointing AI numbers by reaching for a better rollout plan, a tighter implementation timeline, or a vendor with a shinier demo. None of that touches what's actually happening. Employees are filling a silence that leadership left open, and in the absence of real information, they fill it with the most plausible story available: this technology is here to replace them.

That story isn't irrational. These are people who watched automation quietly hollow out entire industries over the past two decades, who learned that "efficiency initiative" is often corporate shorthand for headcount reduction, and who have gotten good at reading between the lines of an optimistic all-hands. When AI shows up without context, without stated boundaries, and without an honest conversation about what it means for their role, silence gets filled with fear by default.

The formula underneath all of this is simple:

Remove any one variable and the equation collapses. Skills without clarity produces employees who can technically use the tool but don't trust what it means for their future. Clarity without skills produces employees who understand the policy but can't do anything useful with it. Trust without either one is just goodwill with nowhere to go.

Why Does Using AI More Often Sometimes Increase Fear Instead of Reducing It?

Employees who use AI daily or multiple times a week are more than twice as likely to say they fear their job will be eliminated within five years, compared to infrequent users, according to Gallup's research this year. As of the first quarter of 2026, roughly 19% of all workers believed their job was somewhat or very likely to be eliminated by AI. Usage is rising and so is displacement fear, and the two are moving together.

CNBC and SurveyMonkey's Q3 2026 survey (published August 19, 2026) shows the same pattern from a different angle. Daily AI users are more optimistic about AI's impact on the job market than occasional users, 35% versus just 7%. But daily users are also still meaningfully pessimistic, at 33%, while occasional users skew almost entirely negative, at 52% pessimistic. On personal job security, daily users net positive, with 27% feeling more secure against 22% feeling less secure. Occasional users net sharply negative, at 3% more secure against 24% less secure. Frequent use doesn't erase fear. It splits people into two camps, and which camp an employee lands in depends on what happens around the tool, not the tool itself.

Gallup identified the actual lever: managers. Employees who strongly agree their organization cares about their wellbeing are six to seven points less likely to report displacement concern. Employees who feel genuinely respected by their manager see that same fear drop by roughly 6.8 points, and the effect concentrates precisely among the heaviest AI users, the group with the most fear to lose. The cost of getting this wrong is measurable too: workers who fear losing their job to AI show lower engagement, higher burnout, lower job satisfaction, and a 4.1 percentage point higher likelihood of actively job hunting. An adoption strategy that raises usage numbers while leaving that fear unaddressed is trading a technology win for a retention problem.

What Did KPMG, Walmart, and Lumen Do Differently?

KPMG, Walmart, Sutherland, and Lumen each built a working model of the "organizational care" and "respect" that Gallup found actually buffers displacement fear, regardless of which side of the platform race they're betting on (see the agentic AI platform war between OpenAI, Amazon, Microsoft, and Google). None of that infrastructure spend pays off without adoption.

KPMG faced the problem most large professional services firms face: a workforce that is highly educated, understandably skeptical, and acutely aware of the compliance risk of getting AI wrong. Instead of deploying a generic AI chatbot and hoping people would figure it out, KPMG built "KymChat," a secure internal interface paired with cross-functional governance and role-specific training on privacy and compliance. The result was organic uptake instead of the shadow IT workarounds that show up whenever employees get a powerful tool with no context.

Walmart and Sutherland both bet that the fastest way to convert skeptics was to put the tool in their hands and let them find the value themselves, backed by bootcamps where employees co-created use cases for their own workflows instead of absorbing a decision made above them. Sutherland made the framing explicit from day one across its 40,000-plus employees: AI was there to eliminate the drudge work, not the people doing it.

Lumen's approach may be the most instructive. Its CEO (chief executive officer) personally modeled Copilot usage rather than mandating it downward, and adoption followed. Employees ended up saving roughly 30 minutes a day, with especially strong results in sales workflows. The bigger outcome was cultural: AI became something the organization did together, not something that happened to people without their input.

What Should Leaders Actually Do to Close the AI Adoption Gap?

Each move below is something a manager can start this quarter, with no new budget line and no new vendor relationship.

1. Publish a plain-language AI policy. Tell people exactly what's expected: which tools are encouraged, what's off-limits, and where the guardrails sit. Most employees aren't afraid of AI itself so much as they're afraid of accidentally doing something wrong with it, and removing that ambiguity is the fastest way to lower the barrier to adoption.

2. Build role-specific training, not generic lunch-and-learns. "AI for account executives" and "AI for demand generation managers" land very differently than "Introduction to Generative AI." People need to see their own workflow in the example before they'll trust it with real work.

3. Run AI pilot squads. Find your early adopters, hand them real workflows to test, and make their wins visible to peers. Adoption becomes credible when it comes from a colleague two desks over, not from a mandate handed down from the top.

How Should You Measure AI Adoption Instead of Licenses and Logins?

Most AI dashboards track licenses deployed, prompts submitted, and tokens consumed. Those are inputs. Here's a scorecard built around outcomes instead:

That third row is new, and it belongs there because Gallup's data made the case for it directly: the same two questions, whether people feel their organization cares and whether they feel respected by their manager, are the strongest predictors of whether higher AI usage produces confidence or fear. If your anxiety scores are flat or climbing while license utilization ticks up, that's a culture signal wearing a technology costume.

Time reinvested only counts as a win if there's somewhere real for it to go. Which automation platform actually absorbs the busywork you're trying to clear is a separate decision on its own terms (see the case for and against Google Antigravity versus n8n).

The Bottom Line

Workplace AI culture is deciding who wins the adoption race this year, not infrastructure. Leaders who model the behavior personally, build real room for safe experimentation, and make it unmistakably clear that AI exists to hand people back time for work that matters end up with organic adoption. Leaders who skip straight to the rollout plan end up with a license count and a workforce quietly polarized between cautious optimism and real fear, exactly what Gallup and CNBC's 2026 data both show happening right now.

Your job isn't to deploy more agents. It's to build a culture where your team treats those agents as the best thing that happened to their Monday morning, not another system quietly grading them from the sidelines. Getting that right also shapes how the market reads your brand's broader AI posture, which is its own kind of exposure (see what a federal ruling on AI governance can teach you about brand risk). This is a leadership job, and it's yours to solve.

If this connects to something you're trying to solve, book a complimentary consulting session. No pitch, just perspective.

Frequently Asked Questions

What is the AI adoption gap?

The AI adoption gap is the space between employees who have access to AI tools and employees who actually use them on a regular basis. IBM's 2026 CEO study puts real numbers on it: 86% of executives believe their employees already have the skills to work with AI, but only 25% of the workforce uses it regularly, a 61-point gap between what leadership assumes and what's actually happening.

How many employees actually use AI at work in 2026?

Gallup's Q1 2026 data, surveyed February 4 to 19, 2026, puts total AI usage among US employed adults at 50%, up from 46% the prior quarter and 40% a year earlier. Daily use sits at 13%, up from just 4% in Q2 2024 (second quarter of 2024), and 28% of employees use AI at least a few times a week.

Does using AI more often increase job security fear instead of reducing it?

Often, yes. Gallup found that employees who use AI daily or multiple times a week are more than twice as likely to fear their job will be eliminated within five years than infrequent users. Roughly 19% of all workers said in Q1 2026 that they believe their job is somewhat or very likely to be eliminated by AI. Usage alone doesn't buy comfort. Whether it builds confidence or fear depends on whether leadership pairs that usage with visible support.

What can leaders actually do to close the AI adoption gap?

Three moves matter most: publish a plain-language AI policy that states what's encouraged, what's off-limits, and where the guardrails sit; build training organized by role instead of generic sessions; and run pilot squads of early adopters whose wins get shared with peers. None of these require new budget or a new vendor.

How should companies measure AI adoption instead of counting licenses and logins?

Track usage depth (the percentage of employees using AI weekly, not just activated), an anxiety index built from pulse survey questions on safety, clarity, and career confidence, manager buffering (whether employees feel their organization cares and feel respected by their manager), and time reinvested (hours reclaimed from low-value tasks and redirected toward customer work or learning). Licenses purchased and workflows automated describe inputs, not outcomes.

Why did KPMG's AI rollout succeed where so many others stall?

KPMG built "KymChat," a secure internal interface paired with cross-functional governance and role-specific training on privacy and compliance, instead of handing employees a generic AI chatbot and hoping they'd figure it out. The result was organic uptake instead of the shadow IT workarounds that show up when employees get a powerful tool with no context.

Does manager behavior actually change how anxious employees feel about AI?

Yes, and Gallup's research makes the effect measurable. Employees who strongly agree their organization cares about their wellbeing are six to seven points less likely to report AI-related displacement concern, and employees who feel respected by their manager see that same fear drop by roughly 6.8 points when they're frequent AI users. The effect concentrates among the heaviest AI users, the group with the most fear to lose.

What does it cost a company when employees fear losing their jobs to AI?

Gallup ties displacement fear to measurable damage: lower engagement, higher burnout, and lower job satisfaction, plus a 4.1 percentage point higher likelihood that a fearful employee is actively job hunting. Raising usage numbers while leaving that fear unaddressed trades a technology win for a retention problem.