Why 49% of Employees Never Use the AI Tools Their Company Already Bought
Start with the number that should worry any company that's bought AI seats this year: as of late 2025, 49% of US employees say they never use AI in their role at all. Not "rarely." Never. Only 26% use it weekly, and just 12% daily. That's from Gallup's own late-2025 tracking survey of 22,368 employed US adults, one of the largest and most methodologically careful workplace-AI datasets available, and it isn't a fringe finding — it's the headline.
The gap isn't access. It's use.
It would be easy to read that number and assume the problem is availability: maybe half of employees just don't have a company AI tool yet. That's not what the data shows. Microsoft and LinkedIn's 2024 Work Trend Index (31,000 people, 31 countries) found that 75% of knowledge workers already use generative AI at work in some form. The gap isn't between "have it" and "don't have it." It's between casual, occasional use and the kind of regular, confident use that actually shows up in output.
Two more numbers from that same Work Trend Index make the shape of the problem clearer:
- 78% of employees who use AI at work are bringing their own tool rather than a company-sanctioned one.
- 60% of leaders admit their organization has no plan to implement AI at all.
Put together: most companies aren't short on AI access. They're short on a coherent internal answer to "which tool, for what, and how do I actually use it well." Employees fill that vacuum by improvising with whatever's free and personal, which is a real productivity story, but not the one a company gets to measure, govern, or build on.
There's also a gap by seniority
Gallup's data breaks this down by role, and the pattern is consistent: leaders use AI at nearly twice the rate of individual contributors (69% vs. 40%, measured as "at least a few times a year"). That's not surprising — leaders usually get first access to new tools, and more discretionary time to experiment. But it means the gap is worst exactly where the largest number of hours are actually spent: the individual-contributor layer of a company, where the 41% of the workweek going to low-value, repetitive work that we've written about separately actually lives.
Buying the tool doesn't close the gap. Training does.
This is the part of the research that should change how a company spends its next AI budget line. BCG's 2026 "AI at Work" study (11,749 workers across 14 markets) found that a clear AI strategy paired with real workflow redesign lifts measured business impact by roughly 25 percentage points — compared to just 5 percentage points from better tools alone. Tools without a plan for how people actually use them move the needle by about a fifth as much as tools with one.
That's the entire argument for training over more software. A company that's already bought Claude, Copilot, or ChatGPT seats has already made the harder decision. The remaining gap — the one keeping 49% of employees at zero — isn't a procurement problem. Our services are built around exactly that layer: a foundational baseline for everyone, then role-specific playbooks matched to what a team has actually licensed, because "we bought the tools" and "the team uses the tools" turned out, in the data, to be two entirely different claims.
If you want to see what this gap is actually costing a specific company, we built an interactive calculator using these same sourced figures — plug in a headcount and see the number.
For why a free vendor course doesn't close this gap on its own, see a free vendor AI course isn't the same thing as training your team. For more on why we think the fix is training specifically, not another platform, see why Aptivate exists.