AI saved you 10 hours. Where did they go?

AI saved you 10 hours. Where did they go?

November 14, 2026•4 min read

AI saved you 10 hours. Where did they go?

Nate B. Jones, wrote this in one of his October 2026 newsletter posts: “Ten hours saved by AI don't become 10 useful hours on their own.”

Here is an example scenario: Someone in the office starts using AI for quotes, customer emails, or the weekly report. Work that took ten hours a week now takes four. A month later the week feels just as full, and nobody can point to where the other six hours went.

It would be easy to manage AI the way we would manage any new tool: buy the subscription, tell the team to use it, and trust that the time saved will show up somewhere. That's a common way to start.

But it often stalls for two reasons. First, saved hours have no assigned job, so they spread into the rest of the week...more email, more meetings, a longer version of the same report. Second, AI creates a new task, which is checking its work. Nate describes it this way: "Cheaper execution increases demand for execution."

A person who used to ask for one analysis can now ask for five, and somebody must read all five before they're used. His line about small businesses is the one to remember: "Forty dollars a month buys a capable model and none of the business process around it."

The subscription is the easy part. Deciding where the hours go is the part worth an owner's attention.

What I'd look at first

1. Put a number on the hours before you count them as saved. Pick a single task where AI already helps. Write down the hours per week it took before and the hours it takes now, measured for one week. Take the time to measure – do not guess or estimate. How you'll know it worked: you can write one line, such as "Quotes: 10 hours/week before, 4 hours/week now."

2. Subtract the checking time. Someone reads the AI's work before it goes to a customer or into the books. Write down who that is and how many minutes per week it takes. For comparison, Nate estimates his own five-step routine takes one person about 20 minutes a week. How you'll know: you have a net number, such as six hours saved minus one hour of review equals five hours net.

3. Give the net hours a job, in writing, before they disappear. Choose one thing the business has put off because nobody had the time: calling prospects to follow up on quotes that didn't close, visiting your top ten customers, or collecting invoices more than 60 days old. Name the person and the day of the week. This shows up as a recurring block on that person's calendar.

4. Point AI at work that somebody already checks. Nate's rule is "start with work you already do and already check." If a person was going to review the work anyway, he or she catches a mistake the same way they always did. If nobody was going to review it, you've added a risk without a way to see it. Schedule the review so it’s not also lost.

5. Look again in 30 days. Check two numbers: the net hours from step two, and the result of the job from step three (callbacks made, visits completed, dollars collected). If the second number is zero, the hours went somewhere else, and step three needs a new owner or a new day of the week.

Left alone, the subscription renews every month, the invoice is the only number anyone sees, and the week stays full. A year from now, the answer to "Is AI paying for itself?" is a feeling instead of a number.

With those two numbers (net hours saved and total hours ‘re-assigned’), you know what AI gave back and what you did with it.

P.S. The 30-day check in step 5 works for any subscription you pay for, not only AI.

Sources

· Nate B. Jones, "Intelligence Is a New Economic Input. We Have to Learn How to Allocate It." (10/25/2026): https://natesnewsletter.substack.com/p/saas-pricing-ai-roi

· Nate B. Jones, "The Human Work AI Agents Create" (8/26/2026): https://natesnewsletter.substack.com/p/managing-ai-agents-at-scale


This article first went out by email in the Data Make You Smile Newsletter. Subscribe to get the next one by email.

Jake Gray

Jake Gray

Jake Gray co-founded a broadband company that he and his partners sold to Charter Communications (Spectrum). He was professor of entrepreneurship at Trinity University for 16 years and has been a CFO and COO. He started Data Make You Smile to help small business owners find where AI and a few simple fixes will save them time and money.

LinkedIn logo icon
Back to Blog