Thinking about AI? Start With the Problem.

By now, you have read/heard about all the wonders that AI can do for your business. If you run a business with 10 people or 200 people, you have heard the noise. Chat bots, AI agents. Copilots. The future of work. Most of it sounds like it was written for someone in Silicon Valley. It does not sound like it was written for someone running a shop, a fleet, a practice, or a services company.

Here is what AI actually means for the rest of us. It is a new way to do work you are already paying people to do slowly, expensively, or badly. That is it. If there is no problem to solve, there is no point talking about AI. But if AI can help you grow revenue, run operations tighter, serve customers better, or cut costs, then it is worth your attention.

Most businesses start with the wrong question. They ask: how do I use AI? That is like asking how do I use electricity. The right question is different. What problem costs me money every single week?

Companies that get real value out of AI follow a consistent pattern. They do not buy AI. They fix one painful workflow. Not a platform. Not a transformation. One thing that hurts. Invoices piling up. Quotes going out late. Nobody knowing which jobs actually made money.

You do not need to treat AI activities as transformational. Transformational mind set and efforts of any scale encounters more resistance, requires more resources and more effort. That approach has challenges. Training. Making those changes stick. The costs involved. Taking incremental steps works better. Start slow. Start with one workflow. Then go from there.

The work stays in the systems you already have. The good projects do not rip anything out. They plug into QuickBooks, the CRM, email, the shared drive. Where the work already lives.

A human still checks the important stuff. The pattern that works is simple. Let the machine do the routine cases straight through. Route anything uncertain to a person. Fewer mistakes. Not fewer humans.

Look for four shapes of pain in your business. Forget use cases.

• Retyping. If someone on your team retypes information from one place to another — invoices into accounting, forms into a system, PDFs into spreadsheets — that is the first thing to automate. It is slow. It is error-prone. It scales horribly. More business just means more typing.

• Waiting for answers. If a simple question takes days and a favor from the one person who knows the spreadsheets — which customers are profitable, which jobs lost money, where is the bottleneck this month — that is a problem AI is good at. Ask in plain English. Get an answer with the receipts attached.

• Things falling through the cracks. Quotes that never get followed up. Calls missed after hours. Complaints that sit in an inbox. These do not show up as a line item. But they quietly cost you customers.

• Doing more without hiring. Every growing small business hits the point where the work grows faster than the team. AI is often the difference between hiring three people and hiring one.

The numbers are coming back from surveys of small and mid-sized businesses. In Thryv’s 2026 survey of 561 small and mid-sized businesses, AI adoption rose to 66 percent from 55 percent a year earlier. Seventy percent said AI increased their revenue over the past year. Fifty-five percent said it reduced costs.

TD Bank’s 2026 survey found 69 percent of small business owners use AI to reduce expenses, up from 39 percent the year before. Customer service, sales, and marketing are the top benefits.

The U.S. Chamber of Commerce found 89 percent of small businesses now use AI in some capacity. But only 14 percent have woven it into their core operations. That is the gap where the efficiencies that can be uncovered with the AI.

Read those numbers with a grain of salt. But, the direction is consistent and businesses can certainly benefit if they point AI at the real problems.

Here is what the surveys also say. Seventy percent of small business owners admit they need more training to use AI effectively. Buying the tool was the easy part. The skill is applying where the value lives.

Businesses that succesfully built and used AI has made it work for them in these areas. A chatbot on your website does not help if nobody taught it your business. A fancy dashboard does not help if nobody trusts the numbers. The winners all did the unglamorous work. Picked one workflow. Cleaned up the data behind it. Trained their people. Measured what changed.

If you take one thing from this post, make it this. Pick one workflow that touches money. Give it 30 days. Write down how long it takes today and what it costs you — in hours, in errors, in lost business. Fix that one thing. Measure again. Then decide whether to do a second one.

And if you look around your business and genuinely cannot name a problem worth solving, skip AI entirely. The technology is the easy part. The problem is the whole game.

References:

1. Thryv, “AI Adoption Continues to Rise but 70 Say They Need More Training to Use It Effectively,” April 2026 https://investor.thryv.com/news/news-details/2026/AI-Adoption-Continues-to-Rise-but-70-Say-They-Need-More-Training-to-Use-It-Effectively/default.aspx

2. DesignLoud, “AI Adoption in Small Business 2026” https://designloud.com/ai-adoption-small-business-2026/