One in five Minnesota businesses now uses AI. Here's what the other four are missing.
Ben McDermott · September 2, 2026 · 3 min read
In May the Federal Reserve Bank of Minneapolis published its read on AI adoption across the Ninth District. Minnesota leads the region. The interesting number isn't the 20 percent who've started. It's why the 80 percent haven't.
The numbers
- 20 percent of Minnesota businesses reported using AI in at least one business function in early 2026, the highest rate in the Ninth District, which averages about 18 percent.
- The Census Bureau's Business Trends and Outlook Survey broadened its question on November 17, 2025, from AI used "in producing goods or services" to AI used in "any of its business functions." Under the old, narrower question, national adoption had reached 10 percent by late 2025, double the mid-2024 rate.
- Large firms with 250 or more employees: 30 percent use AI. Firms with fewer than 20 employees: 17 percent.
- By sector: information and professional services near 40 percent, finance and insurance close behind, healthcare at 20 percent (tripled since 2023), manufacturing at 12 percent (doubled since 2023).
- Roughly 96 percent of AI-using companies reported no change in employment. Where it changed work, AI "primarily supplements employee tasks rather than replacing them."
- Among non-adopters, the most common reason was that AI "did not apply to their operations." About one in five cited safety or knowledge concerns.
The report's author, Erick Garcia Luna, the Fed's director of regional outreach, describes adoption as growing "steadily but unevenly." The unevenness is the story.
The gap is small firms, and the reason is a misunderstanding
A 13-point gap between big companies and small ones would make sense if AI needed a data science team. It doesn't anymore. The things that pay off first in a small business, replying to a lead at 9 p.m., turning a call into a proposal, filing form submissions into a CRM, need a good tool and an afternoon, not a department.
So why do most non-adopters say it doesn't apply to them? Because the picture of AI they've been sold is either a chatbot that gives generic answers or a factory robot. A roofing company owner is right that neither applies. What applies is the pile of repetitive admin that happens between the roofing, and nobody has shown them that's what AI is for.
What "using AI in a business function" looks like at ten employees
Here are three automations we've built with Minnesota clients. None of them required a developer or a data team.
- A weekly note, every Monday, on what moved in search for the business and the three things to do about it. It reads Search Console so the owner doesn't have to.
- After-hours follow-up: a text, call, or email that arrives after close gets a useful answer in the owner's voice within minutes, and the owner finds it logged in the morning.
- A prospecting call, recorded, that becomes a finished proposal PDF in the company's template before the owner is back at the desk.
Each of these is "AI in a business function" by the Census Bureau's definition. Each took one or two working sessions. None of them changed headcount, which matches the Fed's 96 percent.
The safety concern is legitimate and solvable
One in five non-adopters cited safety or knowledge. That's the right instinct, and it deserves a real answer rather than reassurance. The answer is the same one regulated industries use: decide what the tool can see, keep a person reviewing anything that goes to a customer or touches money, and write down who approved what. Our partner Dan Roseman built compliance programs at Coinbase, Circle, Square, and Robinhood, and that's the discipline we apply to a five-person company too.
What happens to the 80 percent
Adoption at the Fed's measured pace means Minnesota crosses 30 percent within a couple of years. The businesses that start now get two years of compounding: each small automation makes the next one easier, because the context files, the accounts, and the habits are already there. The ones that wait will buy the same tools later, at the same price, without the head start.
Sources
- Federal Reserve Bank of Minneapolis, "AI adoption in business grows steadily but unevenly," Erick Garcia Luna, May 29, 2026
- U.S. Census Bureau, Business Trends and Outlook Survey