Who's really using AI?

Everyone says every company is using AI now. We wanted to know if that's true for the companies we build for.

Our customers do $50M to $500M a year in six industries: industrial distribution, manufacturing, construction, field services and trades, trucking and logistics, and professional services. So we pulled the Census Bureau's survey files, including the release that came out on October 22, and read every industry survey we could find. Here's what we found.

01Almost everyone has tried it. Few have built it in.

Ask a mid-sized company if it uses AI and the answer is almost always yes. In RSM's survey of 1,030 mid-market companies in March 2026, 86% said AI was partly or fully built into their operations.²

The Census Bureau asks a stricter question: did the business use AI in any of its functions in the last two weeks? For firms with 100 to 249 employees, 35% said yes. For firms with 250 or more, 44%.¹

Ask whether AI runs a core part of the business, and the number drops again. In RSM's 2025 survey, only a quarter of the companies using AI said it was fully built into their core operations.³ In the Census Bureau's deeper survey last winter, only 6% of all firms used AI to make their products or deliver their services.⁴

These are three different questions with three very different answers. Most companies have AI on someone's laptop. Far fewer have it inside quoting, scheduling, dispatch or billing.

02The physical economy is behind

Across all US businesses, 24.6% told the Census Bureau they had used AI in the two weeks to October 4. In professional services it was 43.8%. In the rest of our industries it was lower: 22.9% in manufacturing, 21.4% in wholesale, 16.1% in construction and 13.5% in transportation and warehousing.¹

03But it's moving fastest there

The Census Bureau changed its AI question last November, so we started the clock there. Since then, the share of manufacturers using AI has gone from about 12% to 22.9%. Wholesale went from about 13% to 21.4%, and construction nearly doubled, from about 8% to 16.1%. Across all businesses, the rate went from about 17% to 24.6%.¹

These industries started further back, and in under a year they've closed much of the gap. Manufacturing is now within two points of the national rate. Construction and trucking are still well behind.

04Size matters, and our customers sit in the middle

Bigger companies use more AI. In the Census data, 24% of firms with one to four employees use it, 35% of firms with 100 to 249 and 44% of firms with 250 or more.¹ Our customers, with roughly 100 to 1,500 employees, sit in those top two groups.

Distribution shows the same pattern more sharply. In the first AI Adoption Index from NAW, the distributors' trade association, 58% of distributors with more than $1 billion in revenue had at least one AI application scaled or fully deployed. Below $50 million, it was 19%.⁵ The largest firms can hire teams to do this work. Most mid-sized companies can't.

05Using AI isn't the same as changing how you work

Even where companies say they use AI, it mostly hasn't changed how they work. In the Census Bureau's survey, 57% of firms using AI used it in three or fewer parts of the business, most often sales and marketing, strategy and IT. Production, supply chain, quality control and distribution were among the least common.⁴

And 64% of the firms using AI had made no changes to support it: no training, no new workflows, no new data or software. Only about 15% had built new workflows around it.⁴ That's what off-the-shelf use looks like.

The industry surveys say the same. Only 10% of manufacturers in Grant Thornton's 2026 survey had AI fully integrated into operations.⁶ In construction, AGC and Sage found 45% of firms using AI for office work, but only 23% for estimating.⁷ In ServiceTitan's survey of more than 1,000 contractors, 52% were actively using AI, but only 11% had high trust in the tools.⁸

06The biggest barrier isn't cost

When the Census Bureau asked firms that don't plan to use AI why not, 65% said it isn't applicable to their business, and 22% said they don't know enough about what it can do. Laws and regulations were among the least common reasons.⁴

Contractors say the same thing in their own words. Their top barriers were understanding how to use AI in their business (44%) and a lack of clear use cases (43%), well ahead of cost (17%).⁸ Distributors put data first: 68% ranked data quality and integration among their top three barriers.⁵

About three in four of the companies we build for haven't put AI into a core process yet.

Most of them have tried the tools. What they haven't had is someone to fit AI to how they quote, dispatch or bill, wire it into the systems they already run, and stay until it works. That's the gap we started Ciao to close.

If you run one of these companies and want to know where AI would fit in yours, book a call. We'll show you a first version on your own work.

Written by

01CEOCIAO Paul Sheridan IN/in/paulfsheridan X@paulfsheridan Founder & CEO01 / 01 CIAOTeam card0101 / 01