Insights
AI is skipping the industries that need it most
If you run a construction company, a distribution business, or a home services outfit, here's a number worth sitting with: only 33% of retail employees, 37% in healthcare, and 38% in manufacturing use AI at work a few times a year or more — versus 76% in technology or information systems, 58% in finance, and 57% in professional services. That's Gallup's Q3 2025 read on U.S. employee AI use, and the gap isn't about which industries are sharper. It's about which industries had knowledge work already sitting in a browser tab when the tools arrived.
The real dividing line
Gallup is careful about how it draws this split, and so should we: it's knowledge-based roles versus frontline jobs, not white-collar versus blue-collar. Plenty of frontline work takes real judgment and real skill — running a job site, servicing equipment, managing a warehouse floor. What it doesn't take is a screen open all day. ChatGPT, Copilot, and the rest of the current generation of tools were built for people who spend their hours in email, spreadsheets, and documents. That's an accident of interface design, not a verdict on which industries are capable of using AI well.
The pattern shows up wherever "frontline" work concentrates:
- Crews and technicians are dispersed across job sites, trucks, and client locations, not desks
- The core systems are legacy, industry-specific software, not the browser-based tools AI plugs into easily
- Decision-makers rarely spend the day inside the workflows their teams actually run
- Adoption, when it happens, is one enthusiast's side project rather than a company process
None of that means operational businesses can't use AI. It means almost none of them have, yet.
The education gap says the same thing a different way
Pew Research Center found the identical divide cut along education instead of industry. Over the past year, the share of workers with at least a bachelor's degree who say some of their work is done with AI rose from 20% to 28%. Among workers with some college or less education, it rose too — but only from 13% to 16%. The gap between the two groups widened from 7 points to 12, and Pew is blunt about why: the overall increase is "mostly driven by" degree-holders.
Same shape, different lens. Whether you slice by industry or by credential, AI adoption is compounding fastest where it already had a foothold — and barely moving where the work happens with your hands, not just your keyboard.
The advantage nobody in your industry has claimed yet
Here's the flip side worth taking seriously: if adoption is this concentrated in tech, finance, and professional services, your actual competitors in construction, home services, distribution, or the trades are almost certainly not using AI well either. Gallup's broader number backs this up — overall U.S. employee AI use rose from 40% to 45% between Q2 and Q3 2025, a fast-moving tide that hasn't reached the operational side of the economy in the same way.
That's a closing window, not a permanent one. But right now, being the first company in your unglamorous industry to actually operationalize AI is an advantage the tech sector spent years ago. Nobody's writing playbooks for a regional HVAC company or a mid-market distributor, which means nobody in your competitive set has one either.
Adoption alone isn't the finish line, though — plenty of pilots launch and go nowhere, for reasons that have nothing to do with the industry running them. That's a separate problem worth understanding before you spend anything. What actually matters here is starting from an honest read of where your operations sit today, not a generic AI strategy borrowed from a software company. That's the order we work in: find out where the real opportunity is, then build the skill inside your team so it doesn't live in one person's head. If you want a fast, unsparing look at where your company actually stands, the AI Ownership Scorecard takes about four minutes.
Source: Gallup, "AI Use at Work Rises" (Q3 2025); Pew Research Center, "About 1 in 5 U.S. workers now use AI in their job, up since last year" (October 6, 2025). For the mechanics of why so many pilots stall regardless of industry, see why AI pilots fail — or start with the AI Opportunity Audit.