More Work Than Workers: The U.S. Mid-Market’s AI Turn

America’s mid-market grew its revenue by 11 percent over the past year, while its headcount grew more slowly than at any time since the pandemic. That gap is, for the most part, not a layoff story. It is what artificial intelligence looks like in companies that have more work than people to do it – and it is about to change what those companies are worth when their owners sell.

The figures come from the National Center for the Middle Market, which surveyed 1,000 executives of American companies with annual revenues between $10 million and $1 billion in June. Eighty-two percent of them reported higher revenue than a year earlier, nearly half by double digits. Employment grew too, by 7.2 percent – healthy by any standard, but the slowest pace the center has measured since the pandemic. And nine out of ten of these companies now say they use AI.

Put those three numbers side by side and a pattern emerges that fits neither of the stories usually told about AI. It is not the story of machines replacing workers: more than half of these companies are still adding staff, and 59 percent expect to keep hiring through mid-2027. Nor is it the story of a technology still waiting to prove itself. It is a story about capacity – about companies that have found a way to take on more work than their payroll alone would allow.

A third of the economy, rarely in the headlines

The middle market is easy to overlook because it sits between two better-told stories. Small business has its own federal agency, its own loan programs and its own political vocabulary. Large corporations have analysts, earnings calls and AI budgets measured in billions. In between are roughly 200,000 companies – about three percent of all American businesses – that account for around a third of private-sector GDP and employment, some 48 million jobs.

They are regional manufacturers, distributors, engineering firms, logistics providers, health care groups and software houses, many of them family-owned or backed by private equity. They are big enough to run real processes – payroll for hundreds of people, supply chains across several states – and small enough that the CFO still knows which spreadsheet holds the truth. That combination makes them the part of the economy where AI either becomes ordinary operations or stays a slide in a board deck.

The labor math has changed

The pressure behind all this is demographic, and it is not going away. The Bureau of Labor Statistics projects that the U.S. labor force will grow by just 0.3 percent a year between 2024 and 2034. Labor force participation is expected to fall from 62.6 percent to 61.1 percent over the same decade as the last baby boomers age out of work. According to an analysis of the BLS projections by Indeed Hiring Lab, that decline alone means roughly 4.3 million fewer workers in 2034 than steady participation rates would have produced.

For a mid-market company, that arithmetic is harsher than the national average suggests. It recruits from the same shrinking pool as the large corporations, usually without their pay scales, brand names or training programs, and often in regions where the next qualified applicant lives an hour’s drive away. When a controller, a dispatcher or a quality engineer retires, the position does not stay open for a few weeks. It stays open.

The companies using AI are the ones hiring

This is where the data contradicts the most common fear about AI. In April, the National Center for the Middle Market compared the companies in its year-end survey that use AI with those that do not. Eighty-seven percent of the AI users had grown their revenue over the previous year, compared with 66 percent of the others; their average growth was 12.9 percent, against 5.8 percent. Six in ten AI users had increased their workforce, compared with 39 percent of non-users – and the AI users expected to grow their headcount twice as fast in the coming year, by 10 percent instead of 5.

Correlation is not causation: fast-growing companies are also the ones with the money and the appetite to try new tools. But the pattern cuts against the idea that AI in this segment is mainly a way to cut staff. What it suggests instead is that AI absorbs the work for which there are no people – the invoices that would otherwise pile up, the quotes that would go out a week late, the reports nobody had time to write – and frees the people a company does have for the work only they can do.

Executives expect the shift to go further. In RSM’s Middle Market AI Survey 2026, 88 percent of respondents said the composition of their workforce would look fundamentally different within two to three years, and 91 percent expect people and AI systems to work together as integrated teams. Eighty-five percent also conceded that their leadership is more enthusiastic about AI than their employees are – a gap no software will close.

From pilots to plumbing

By most measures, the experimenting phase is over. RSM found that 86 percent of mid-market companies have partially or fully integrated AI into their operations, that 73 percent use generative AI and that 58 percent plan to invest $1 million or more in AI this fiscal year. Ninety-seven percent say they are satisfied with what they have spent so far.

What holds them back now is less glamorous. Asked what stands in the way of scaling AI, 53 percent named data quality and 47 percent integration with existing systems. KeyBank’s Middle Market Sentiment Survey, which polled 750 owners and senior executives at the end of 2025, points the same way: three quarters of the companies plan to automate employee tasks, but only six percent say they have reached the state of AI use they are aiming for.

That is the real work of the next few years, and most of it is plumbing: getting customer, order and inventory data into one consistent place, connecting the ERP to the tools around it, writing down the processes that so far have lived in the heads of a few long-serving employees. None of it makes headlines. All of it decides whether an AI system can be trusted with a task or only with a suggestion.

Paid for out of the business

How mid-market companies pay for all this is telling. In the KeyBank survey, 53 percent named AI and technology expansion as a priority for their investments, ahead of equipment, hiring and new products. Yet 57 percent were content with the credit lines they already have; only 43 percent were adding capacity through debt or equity. KeyBank’s reading: many companies fund efficiency from within first and bring in outside capital only to scale what has already proven itself.

That is a sensible order of operations for a technology whose benefits show up in the income statement long before they show up anywhere a lender could take as collateral. It also means that the pace of AI adoption in the middle market is set by cash flow – one more reason why the companies that are already growing are likely to pull further ahead.

The succession test

The most consequential effect may show up when these companies change hands. The McKinsey Institute for Economic Mobility estimates that about six million small and midsize American businesses will face an ownership transition by 2035 as their owners retire. More than a million of them are viable candidates for a sale, representing up to $5 trillion in enterprise value.

Buyers do not pay for a founder’s memory. They pay for a business that keeps running when the founder leaves – processes that are documented, data that can be audited, customer relationships that live in a system rather than an address book. Among smaller owner-run firms, the gap is wide: in a 2026 survey for Zelle, only 29 percent of business owners over 50 described their operations as modernized, 60 percent had no formal succession plan, and 67 percent of the younger would-be buyers said outdated payment systems alone could derail a deal.

The middle market is better prepared than Main Street, but the logic is the same, and private equity buyers apply it with spreadsheets. A company whose order handling, bookkeeping and customer service already run on reliable data and well-defined workflows – with AI doing part of the work – is easier to diligence, easier to finance and easier to run for a new owner. In the coming decade, that may become one of the most tangible ways AI turns into value: not as a line in the budget, but as part of the price.

Three questions for the year ahead

The first is whether hiring holds up. If revenue keeps growing faster than headcount and the AI users keep adding staff, the capacity story is confirmed. If employment growth falls further while revenue holds, the replacement story will gain ground – and with it the resistance that RSM’s enthusiasm gap already hints at.

The second is data. The companies that clean up their master data and connect their systems in the next twelve months will be able to give AI real responsibility. Those that do not will keep running pilots.

The third is the deal market. As the wave of ownership transfers gets under way, due diligence will increasingly ask how a company’s processes work without its key people. The answers will show which companies treated AI as a tool and which made it part of the business.

America’s mid-market has rarely been the first to adopt a technology, and it has rarely needed to be. It adopts what makes the business more durable. With more work than workers for the foreseeable future, AI has quietly moved into that category.

---

Sources and further reading

• National Center for the Middle Market, Mid-Year 2026 Middle Market Indicator: U.S. middle market sustains double-digit growth while AI fuels a shift toward productivity, July 29, 2026
• National Center for the Middle Market: AI adoption in the middle market isn’t about experimentation, it’s about performance, April 30, 2026
• National Center for the Middle Market: Middle Market 101
• RSM: The middle market has embraced AI. Now comes the hard part., July 21, 2026
• RSM Middle Market AI Survey 2026: The building blocks of AI adoption
• KeyBank: Middle Market Sentiment Report: Why leaders are starting 2026 on the front foot, February 2026
• U.S. Bureau of Labor Statistics, Career Outlook: The economy and labor force, projected 2024–34, September 2025
• Indeed Hiring Lab: Why labor force participation is projected to fall through 2034, April 7, 2026
• McKinsey Institute for Economic Mobility: The Great Ownership Transfer: A new era of business stewardship, February 26, 2026
• Zelle Small Business Pulse Report 2026: Boomers built Main Street. Now they’re ready to sell – but Gen Z and Millennials aren’t buying, PR Newswire