6 min read
86% of small business owners in one new survey now use AI, but only 15% say it is fully built into how their business runs. Those numbers come from a Babson College survey of Goldman Sachs 10,000 Small Businesses graduates, shared at the launch of Goldman’s Small Business AI Institute and reported by Fortune. The gap between “I use ChatGPT sometimes” and “AI is part of how we work” is where most owners are stuck.
This illustrative story shows one way across that gap. Marco is a composite character, built from common patterns among solo service businesses, not a real person, and his numbers are realistic estimates rather than reported results. But the steps he takes are ones you can copy this week, whatever your trade.
Meet Marco, the occasional AI user
Marco runs a one-person landscaping and garden maintenance business. He has about 35 regular clients, picks up a few design and install jobs each season, and does his admin at the kitchen table after dark.
He was not an AI skeptic. Like most of the owners in that survey, he used a chatbot here and there: once to write a Facebook post, once to settle on wording for a price increase letter, once to ask about a plant disease. Each time it helped. But each time he started from a blank chat, retyped his background, and forgot what had worked. AI was a tool he visited, not a system he relied on.
His real problem was the same every week. Every job created a trail of small tasks: a quote for new work, a follow-up after the visit, a note about what to do next season, a reminder about a payment. Some of those slipped. A few clients drifted to competitors who simply responded faster.
The shift: pick one workflow, not ten tools
The mistake Marco almost made was trying to “do AI” everywhere at once: marketing, bookkeeping, scheduling, social media. That is how owners end up with five free trials and no habit.
Instead, he asked one question: which single process touches every job I do? For him, it was the moment after a site visit, when he has to turn what he saw into a quote, a follow-up, and notes for later. Every client, every week. If that one moment ran smoothly, everything downstream would too.
That is the lesson worth underlining. Moving from dabbling to built in does not mean using AI for more things. It means using it for the same thing, the same way, every single time.
What Marco built, step by step
1. A one-page facts sheet
He wrote down everything an assistant would need to know: his services, his price ranges, his service area, his payment terms, his scheduling rules, and a few phrases he always uses with clients. He saved it as standing instructions in a project in his AI tool so he never had to retype it. This alone made every answer more accurate, because the AI stopped guessing about his business.
2. A voice note after every visit
Before leaving each property, Marco records a two-minute voice memo on his phone: what the client asked for, what he noticed, rough measurements, and anything to remember. Talking in the truck is faster than typing at night, and nothing gets forgotten between the job and the desk.
3. One saved prompt that does three jobs
Each evening he pastes the transcribed memos into his AI project with the same prompt:
“For each visit below, write: (1) a short quote with line items using my price ranges, (2) a friendly follow-up email in my tone, and (3) a three-line note for the client’s file about what to check next season. Flag anything where you are unsure of a price instead of guessing.”
That last sentence matters. It tells the AI that “I am not sure” is an acceptable answer, which cuts down on invented numbers. Our guide to turning a discovery call into a proposal uses the same principle for consultants.
4. A five-minute human review
Marco reads every quote before it goes out, checks each price against his own list, and fixes anything that sounds unlike him. AI drafts, Marco decides. Nothing is sent automatically.
5. A weekly reminder pass
On Friday afternoons, he asks the same project: “Based on this week’s notes, which clients have I quoted but not heard back from?” It gives him a short list, and he sends a quick nudge to each. That is where the lost jobs used to leak out.
What changed after six weeks
In this illustrative scenario, the changes are modest but they compound, which is exactly why they matter:
- Quotes go out the same evening instead of two or three days later. Clients notice speed more than polish.
- Evening admin drops from roughly an hour a night to about 25 minutes, because he edits drafts instead of writing from scratch.
- Fewer quotes go unanswered, thanks to the Friday follow-up list.
- Next season is already planned, because every client file has notes on what to check in spring.
Notice what Marco did not do. He did not buy a premium AI plan, hire a consultant, or connect ten apps. He used the same chatbot he already had. The difference was structure: a facts sheet, a capture habit, one saved prompt, and a review step. That fits what the survey suggests too: most owners already have access to capable tools, and what they say they need is more training and support.
How to find your own “one workflow”
If you want to follow Marco’s path, start by listing the moments that happen for every customer. For most solo businesses, it is one of these:
- After a first conversation (consultants, designers, coaches): call notes become a proposal and a follow-up.
- After a site visit or photos (trades and home services): observations become a quote. Our house painter story shows the photo version.
- After a sale or appointment (shops, salons, studios): a thank you, a review request, and a rebooking reminder.
- After a review comes in (any local business): a thoughtful reply, as in our guide to answering every Google review in five minutes a week.
Pick the one that happens most often and costs you the most when it slips. Then build the same four pieces Marco did: facts sheet, capture habit, saved prompt, human review. Free tools work fine for this. Google, for example, has added a small business hub inside Gemini, and ChatGPT and Claude both let you save project instructions.
Your first step this week
Do not try to build the whole system at once. This week, just write your one-page facts sheet and save it in your AI tool. Next week, add the saved prompt for your one workflow. You will feel the difference by the third job, and you will have moved from the 86% who use AI to the much smaller group who actually rely on it.
Which moment happens for every one of your customers? Share it in the comments, and we will suggest a starter prompt for it.
