How small and mid-sized businesses are growing with AI: what the 2025 data shows

The US Chamber of Commerce's 2025 Empowering Small Business report found that 58% of small businesses use generative AI, up from 40% in 2024 and 23% in 2023 ^[1]. Almost 60% say they use AI somewhere in their operations.

Larger companies moved the same way. McKinsey's 2025 global survey found 88% of organisations regularly use AI in at least one business function, up from 78% a year earlier ^[2]. If a competitor is not using AI yet, they are now the exception.

The fear that AI only means fewer jobs does not match what small businesses report. In the Chamber survey, 82% of small businesses using AI increased their workforce over the past year ^[1].

That is a correlation, not proof that AI caused the hiring: growing businesses are also the ones most likely to try new tools. What it does show is that AI adoption and growth are happening in the same companies, and that owners are using AI to take on more work rather than only to cut cost.

McKinsey's numbers show the gap. While most organisations use AI, only 39% report any impact on EBIT at the enterprise level, and only about a third say they have begun scaling AI across the business ^[2].

McKinsey calls the companies that attribute 5% or more of EBIT to AI, and report significant value from it, "AI high performers". They are about 6% of respondents ^[2].

The clearest difference is how AI is put to work. McKinsey found high performers are nearly three times as likely as others to fundamentally redesign their workflows when they deploy AI ^[2]. They pick a process, change who does what and where AI sits in it, and measure the result.

For a small or mid-sized business that translates into a short list:

• Start with one process that costs money every week: lead response, quoting, invoicing, scheduling or support.
• Redraw the steps with AI doing the repetitive part and a person approving what matters.
• Connect it to the systems the work already lives in: CRM, email, calendar, accounting.
• Agree the number you will judge it on before you build, such as hours saved, response time or conversion rate.

Found this useful?

Share the key finding on X.