How much time can AI automation save a small business?

Claims that AI will save you “20 hours a week” mostly come from people selling it. The independent numbers are more modest and more useful, because they tell you what to expect and where the real gains hide. Below are the benchmarks, then four ordinary businesses and the workflow each should automate first.

What independent research says

Researchers at the Federal Reserve Bank of St. Louis surveyed U.S. workers in November 2024. Among those who had used generative AI in the previous week, average time savings were 5.4% of work hours, about 2.2 hours in a 40-hour week.^[1] Frequency mattered: 33.5% of daily users said it saved them four hours or more, compared with 11.5% of people who used it on just one day that week.^[1]

Thomson Reuters’ 2025 Future of Professionals report surveyed 2,275 professionals across legal, tax, accounting and related fields. Respondents predicted AI would save them an average of five hours per week within the next year, up from four hours predicted in 2024.^[2] That is a forecast, not a measurement, but it comes from people doing document-heavy work every day.

Adoption is no longer the obstacle. Almost 60% of U.S. small businesses say they use AI, according to the U.S. Chamber of Commerce’s 2025 Empowering Small Business report.^[3] The gap is between opening a chatbot now and then and automating a workflow that runs every day.

Chat assistants and automations save time differently

A chat assistant helps when you sit down to do a task: drafting an email, summarising a document, rewording a quote. It makes the task faster. An automation runs without you: an enquiry arrives, and the system replies, logs it and offers booking times. It removes the task.

The Fed’s figures mostly reflect the first kind. The second kind is where a small business recovers hours every week, because the work stops landing on someone’s desk at all.

Four ordinary businesses, four workflows

The examples below apply the published benchmarks to typical small businesses. The time estimates are worked examples with the assumptions stated, not client results. Swap in your own numbers.

A plumbing or HVAC company: instant lead response

The problem here is speed. In an audit of 2,241 U.S. companies published in Harvard Business Review, firms that tried to contact an online lead within an hour were nearly seven times as likely to qualify it as firms that waited even an hour longer. Among firms that responded at all, the average response time was 42 hours.^[4] The study dates from 2011, and it remains the most-cited benchmark on lead response.

In a trades business, the owner is usually on a job when the web form arrives. The automation: every form, missed call and after-hours message gets an immediate text or email with a booking link, the lead is logged in one place, and the office gets a summary to follow up.

Worked example: 20 enquiries a week at 8 minutes each to read, reply and log is about 2.7 hours. Automating the first reply and the logging, while a person still makes the call, could save around half of that. The larger gain is that nobody waits 42 hours.

A dental clinic: reminders and confirmations

An empty chair is time a clinic can’t resell. A meta-analysis in BMJ Open found patients who received text notifications were 23% more likely to attend. No-show rates were 15% with notifications and 21% without.^[5]

The automation: reminders at set intervals, confirmations captured by reply, and a cancellation that automatically offers the slot to a waitlist.

Worked example: 150 appointments a week at 2 minutes per manual reminder call is 5 hours of front-desk time. If automated texts confirm most of them and staff only call the few who don’t reply, roughly 4 of those hours come back.

A hair salon: booking and rebooking

Salons lose time to bookings by phone and direct message, and revenue to clients who forget to rebook. The automation: online booking that syncs with the stylists’ calendars, an automatic confirmation and reminder, and a rebooking prompt sent at each client’s usual interval.

Worked example: 60 bookings a week handled by phone or message at 4 minutes each is 4 hours. Moving most of them to self-booking frees the bulk of that time, and nobody has to put the scissors down to answer the phone.

An accounting or bookkeeping firm: document intake

Tax season is a document chase. The automation: a client form or portal that requests the specific documents each client owes, sends reminders until they arrive, and sorts them by type. AI can draft first-pass summaries and routine client emails for a person to review. This is the document-heavy work behind the five hours a week Thomson Reuters’ respondents expect to save.^[2]

How to estimate your own savings

One formula covers it: times per week × minutes each × share that can be automated ÷ 60 = hours saved per week.

• Log one ordinary week. Note every repeated task and roughly how long it takes.
• Circle the tasks that follow rules. Same trigger, same steps, same result: reminders, confirmations, intake forms, status updates.
• Estimate the automatable share honestly. A reminder is almost fully automatable. A complaint is not.
• Start with the biggest number. Automate one workflow, measure it for 30 days, then pick the next.

What not to automate

Anything that needs judgement: complaints, pricing exceptions, bad news, an upset customer. Automate the routing, not the response. And for anything AI writes that goes to a customer, have a person review it until the output has earned your trust. A wrong answer sent instantly is not a time saving.