Most small businesses have already tried AI. In a 2026 Goldman Sachs survey of small business owners, 76% said they use it, but only 14% said it’s fully part of how they operate. The gap isn’t interest. It’s knowing where to start.
The same survey found 73% of owners would benefit from more training and help putting AI to work (Goldman Sachs 10,000 Small Businesses Voices, 2026). Here’s the approach we use with Central New York businesses.
- 01Pick one taskRepetitive, with no judgment call each time
- 02Set simple rulesApproved tools, what stays out, who reviews
- 03Try it for two weeksTime the task before and after
- 04Keep, change or dropThen pick the next task
Start with one task, not a strategy
Pick one repetitive task that eats hours every week and doesn’t need a judgment call every time. Good first candidates:
- Drafting replies to the same customer questions
- Turning meeting notes into a summary and a task list
- Writing first drafts of quotes, proposals or job descriptions
- Summarizing long documents, contracts or reports
- Cleaning up and reformatting spreadsheets
One task is enough. If it works, you’ll know exactly what to try next. If it doesn’t, you’ve lost a week, not a year.
Use the tools you already pay for
You may not need anything new. Microsoft 365 and Google Workspace both include AI assistants on many business plans, and CRMs, accounting software and email platforms are adding AI features every month. Check what you already have before buying another subscription.
If you do add a standalone assistant like ChatGPT, Claude or Gemini, use a business plan rather than personal accounts, so your company controls the accounts and the data settings.
Set simple rules before you scale
A one-page policy prevents most problems. Decide which tools are approved, what information never goes into them, and that a person always reviews anything AI writes before it reaches a customer. We cover this in a simple AI policy for small business teams.
Measure the time saved
Before you start, write down roughly how long the task takes today. After two weeks with AI, check again. Real numbers make the next decision easy, and they help your team see the point.
When to bring in help
If your team is stuck after the first experiment, or the task involves customer data, several tools or your own documents, that’s when outside help pays off. A short working session usually finds two or three uses worth keeping. Learn about our AI consulting.
Founder and lead developer of DigitalRocket in Syracuse. Jon has spent more than 20 years building websites and software, including over a decade at Syracuse advertising agencies, and now helps Central New York businesses put AI to work, connect their tools with custom software, and run fast, secure websites.