Two-Thirds of Small Businesses Now Run AI. Seven in Ten Say They Still Need Training to Run It Well.

Read those two numbers together and the picture gets clear. The tools are on. The skill to run them well is the part still being built.

Comfortable is not the same as effective

Grant Freeman, President of Thryv, put it plainly in the release. "Most small businesses are comfortable using AI. But being comfortable isn't the same as being effective." He went further: "Closing the gap between AI adoption and capability will be critical for Main Street success."

"Most small businesses are comfortable using AI. But being comfortable isn't the same as being effective."

The survey backs the comfort claim. 86% of respondents said they are comfortable using AI. That is a high number, and it is the easy part, because comfort is what gets a tool adopted. It is not what gets a return out of it.

Here is where a leader should pay attention. When 86% feel comfortable and 70% still want more training, those two groups overlap heavily. People are using the tools daily and know, at the same time, that they are not yet getting everything the tools can give. That honesty is a good sign. It means the market is past the novelty and into the work.

The results are real, which raises the stakes

This is not a story about AI failing to deliver. The Thryv numbers point the other way.

• 92% said AI saves them time.
• 79% expect between 11 and 60 hours of time savings per month.
• 70% said AI increased their revenue.
• 55% reported reduced costs.
• 61% estimated $500 to $2,000 in monthly savings.
• 81% said AI made them more strategic.

Those are self-reported figures from a marketing-platform vendor's survey, so treat them as what business owners believe about their own results rather than audited outcomes. Still, the direction holds. Owners are spending on this. 53% now put at least $100 a month toward AI, and 33% are spending more than they were a year ago.

One figure should catch the eye of anyone thinking about headcount: 46% said they would choose AI over hiring a new employee. That is a labor decision, a budget decision, and a training decision all at once. A business owner who picks a tool over a hire is betting that someone on the team can operate that tool at a level that replaces real work. If the skill is not there, the bet does not pay off.

So the better the results get, the more the training shortfall costs. A tool that already saves time and lifts revenue is a tool worth learning to run at full strength.

A second survey points the same direction

The Thryv finding does not stand alone. Freeman referenced it himself: "The Thryv findings align with a recent Goldman Sachs survey in which 73% of small businesses said they need additional training."

Goldman Sachs ran its own numbers through its 10,000 Small Businesses program. In that survey of 1,256 participants, fielded January 27 through February 4, 2026 by Babson College and David Binder Research, 76% said they currently use AI. And 73% said they would benefit from additional access to training and implementation resources.

Two independent surveys, different samples, different months. Both point the same way. Adoption is high. The demand is for skill and support, not for another product.

Where the shortfall actually lives

There is a useful way to think about the distance between "we use AI" and "we get results from AI." It is a proficiency question, and proficiency can be measured.

That is the whole reason The 7 Levels of AI Proficiency exists. Turning a tool on is the first step. Learning to ask it the right way, to check its work, to hand it real tasks with judgment, and eventually to design workflows around it, those are distinct skills that build on each other. A team can be fully adopted at the tool layer and still be early on the capability climb. The 86%-comfortable, 70%-still-need-training split is that exact situation, described in a founder's own words.

The encouraging part for any leader reading this: you are not late, and you are not alone. Two-thirds of the market is already in. The next move is the one that separates the businesses seeing $500 to $2,000 a month from the ones still poking at the tool. That move is skill, and skill is trainable.

What a leader can do with this

Training is the lever both surveys point to, and it is a decision an owner controls this quarter.

Start by naming the workflow before naming the tool. The businesses getting real savings tend to have one clear job the AI does well, not a scattered pile of experiments. Pick the process that eats the most hours, and build capability there first.

Then measure where your team actually stands, honestly. Comfort surveys will tell you people like the tool. A proficiency read tells you whether they can run it well enough to change a P&L. Those are different questions, and only the second one shows up in the numbers Thryv and Goldman Sachs are reporting.

Related reading: Level 5: The Captain (Design Thinker).

Sources

1. AI Adoption Continues to Rise, but 70% Say They Need More Training to Use It Effectively (Thryv, July 9, 2026)
2. Small Businesses Embrace AI but Need Training and Support (Goldman Sachs, 2026)

Harrison Painter

Executive AI Advisor. Founder, LaunchReady.ai and AI Law Tracker.

Harrison is an Indiana AI Advisor who helps business owners and executives get their time back by building AI systems that run the work for them. Nearly 20 years in business and author of You Have Already Been Replaced by AI. Creator of The 7 Levels of AI Proficiency.

Connect on LinkedIn