Your Team Already Uses AI. Nobody Taught Them How.
The question of whether small businesses will adopt AI is settled. A Thryv survey published this month puts AI adoption among US small businesses at 66 percent, up from 55 percent a year ago. Pax8’s new research tells the same story, with two in three SMBs saying AI makes them feel more competitive, not less.
Here is the number that should get your attention instead: 70 percent of those same owners admit they need more training to use AI effectively. And when asked where they currently learn, 57 percent said YouTube and social media.
Read those together and the picture is clear. Your team is already using AI, probably daily, and almost nobody taught them how. The competitive gap between small businesses is no longer about who has AI. Everyone has it, or will within a year. The gap is between teams that use it well and teams that use it badly while believing they use it well.
What untrained AI use actually costs
Untrained use looks productive from a distance. Everyone is busy, drafts appear faster, tasks get ticked off. The costs hide underneath.
There is rework: output that reads well but is wrong in ways the writer could not spot, so someone senior quietly fixes it. There is inconsistency: five people prompting five different tools five different ways, producing customer emails in five different voices. And there is risk: the employee who pastes a customer list or a contract into a free chatbot because nobody ever said not to. None of this shows up in an adoption statistic. All of it shows up in your margins eventually.
The wider industry has noticed where the real problem sits. This week TechCrunch reported that Anthropic and Blackstone are betting the next trillion dollar AI business is implementation, not models. When the companies building the models say the money is in helping people use them properly, that tells a small business owner exactly where the leverage is.
Training that fits an SME budget
The good news is that closing this gap does not require an enterprise training program or a consultant embedded for six months. It requires a few deliberate moves.
Start with tasks, not tools. List the five tasks where your team already uses AI most: quotes, support replies, reports, whatever they are. Improve those five before touching anything new. Generic “learn AI” training fails because it trains nothing in particular.
Write one page of rules. Which tools are approved, what data must never be pasted into them, and who checks AI-assisted output before it reaches a customer. One page, written this week, beats a policy project that ships next quarter.
Make it social. A 30 minute session every two weeks where one person shows how they actually use AI on real work will outperform any course. Keep a shared document of prompts that worked. This costs nothing and compounds fast.
Give it an owner. Not a new hire, just one existing person who cares, keeps the prompt library alive, and evaluates one new tool a month so everyone else can ignore the noise.
The point
The surveys say adoption is done and competence is not. That is good news if you move on it, because your competitors are reading the same headlines and concluding the job is finished. Two hours a month of structured practice will put your ten person team ahead of most fifty person teams. The tools are the cheap part. They always were.