AI helps beginners most
There's a question I hear from almost every experienced business owner when AI comes up:
"I've been running this business for fifteen years. How can AI really help me?"
It's a fair question. And the research has a surprising answer. AI may not make you dramatically better at what you already do well. But it might make everyone around you a lot more like you.
5,000 customer support agents
In 2023, economists Erik Brynjolfsson, Danielle Li, and Lindsey Raymond published one of the first large studies of generative AI in a real workplace. They followed 5,179 customer support agents at a software company as an AI assistant was gradually rolled out to them. The assistant followed customer chats and suggested responses in real time. Agents could use the suggestions or ignore them.
On average, the agents with access to AI resolved about 14% more customer issues per hour. But the average hides the real story. The newest and least-skilled agents improved by about 34%. The most experienced, highest-skilled agents? Their improvement was minimal.
The AI didn't help everyone equally. It helped the people who had the most to learn.
Why this happens
The researchers found evidence for a fascinating explanation. The AI had been trained on data from the company's past customer conversations, including the ones handled by its best agents. So when it suggested a response, it was often suggesting what a top performer would have said.
In other words, the AI was quietly spreading the best agents' know-how to everyone else. New agents got a kind of real-time coaching from the company's top people, without those people having to do anything.
The study also found that newer agents with AI access moved up the learning curve faster. And customers were happier, and the agents were less likely to quit. A separate 2023 experiment by MIT's Shakked Noy and Whitney Zhang found a similar pattern with professional writing tasks: people who started out as weaker writers gained the most from using ChatGPT, which narrowed the gap between them and the strongest writers.
What this means for a small business
If you run a small business, you probably are the top performer. You're the one who knows how to handle the tricky client, write the proposal that closes, and explain your service in a way that clicks. The bottleneck is usually that nobody else can do it quite like you.
That's where this research gets exciting. AI's biggest strength may not be doing your job faster. It may be helping your team, or your new hire, or your assistant, do their jobs more like you would.
The real opportunity isn't making the expert faster. It's making your expertise available to everyone else.
How to use this
1. Capture how you do things. The AI in the study worked because it had examples from top performers. Yours needs examples too. Save your best proposals, your favorite client emails, your answers to common questions, your approach to handling objections.
2. Build AI tools around your examples. Give your AI assistant those examples and ask it to help your team write in your style or follow your approach. A new hire with your best 20 client emails and an AI assistant is going to ramp up a lot faster than one starting from scratch.
3. Use it for onboarding. New team members often take months to learn "how we do things here." AI tools built on your real examples can shorten that.
4. Don't write it off because it doesn't impress you. If you try AI on your own specialty and think, "This isn't as good as me," you might be right. But ask a different question: Is it better than what your newest team member would produce alone?
5. Keep reviewing. Even in the study, agents could choose to accept or reject the suggestions. Good judgment still matters. AI can give a beginner a head start, but someone experienced should still be checking the work.
Recently I partnered with a company to help install AI systems in their business. One of the biggest issues they were struggling with was with their sales team. Not only did they have a moderate churn rate for their sales team, but the newest members were consistently struggling to get their close numbers up. As part of the system setup, I had the sale manager interview his highest performers about their best practices, pulled their toughest closing assets (responses to texts, emails, and call transcripts) and compile everything into project files for the “sales team” project we created. From there we were able to build a skill that handles sales objections using the best practices and real examples of his highest performers and pass it along to new members of his team. Over the following 8 weeks, not only did closing rate sky rocket, but he didn’t need to hire a new sales team member for the first month in the last year.
Your expertise, multiplied
The experienced business owner's instinct about AI isn't wrong. It may not dramatically change how you work. But if you're the bottleneck in your business, the one everyone comes to because you know best, AI might be the tool that finally lets your knowledge scale.
You spent fifteen years getting good at this. Now you can share it.
If you want help figuring out where AI fits in your business, let's talk it through.
Studies to link: Brynjolfsson, Li & Raymond, "Generative AI at Work" · Noy & Zhang (2023), Science
