Last weekend, I was in the San Francisco Bay Area for a client meeting and also had the chance to watch my daughter play college soccer. The University of Louisville upset #6-ranked Stanford and then tied Cal. I am blaming the euphoria from the Stanford win for postponing last week’s blog.

Somewhere between work and the games, Abby and I had a small window to explore downtown San Francisco together. She had never experienced downtown San Francisco before, so we decided to make the most of it.

On our way to get a closer view of Alcatraz from Pier 39, we rented Lime scooters and, thankfully, survived the two-mile ride along the Embarcadero. On the way back, we decided to try something new.

I downloaded the Waymo app and summoned a vehicle.

That seemed like a great idea right up until we climbed into the driverless car.

The previous passenger apparently spoke Mandarin. The screen was in Mandarin, the instructions were in Mandarin, and the voice speaking to us was in Mandarin. We had absolutely no idea what our driverless car was trying to tell us.

After nearly 10 trips to China over the years, I could remember Ni Hao and Xie Xie, “hello” and “thank you.” That pretty much exhausted my Mandarin vocabulary.

Not very helpful.

My daughter finally pulled out her phone and used a translation app just to confirm that, yes, the car really was speaking Mandarin.

There was something beautifully ironic about using one piece of advanced technology to figure out what another piece of advanced technology was trying to tell us.

So we called support. Their solution was straightforward: get out of the car, and they would send us another one.

Fair enough.

The second Waymo arrived, and we climbed in. We were only about two miles from where we needed to go, and traffic was not particularly bad. In my mind, this seemed simple: just head back down the Embarcadero.

Instead, the app told us the trip would take roughly 45 minutes, travel down about 15 different streets, and cost approximately $45.

For two miles!!

At that point, I started wondering whether we had accidentally selected the scenic tour package.

The bigger issue was that my daughter had to get back for a coaches’ meeting, and being late was not really an option. We called support again and asked why our incredibly sophisticated autonomous vehicle appeared determined to take the longest possible route through San Francisco.

We were told the more direct roads were too busy.

They weren’t.

They explained the routing logic and offered us a $10 credit. I appreciated the gesture, although $10 did not quite solve the problem of paying $45 to travel two miles at approximately the speed of a determined pedestrian.

So, once again, we grabbed Lime scooters and raced back to my car before the two-hour parking meter expired.

We almost made it.

The parking enforcement officer was getting ready to ticket me as we arrived about 20 minutes late. We jumped into the car, raced across the Bay Bridge, and managed to get Abby back in time for her meeting.

Whew!

Somewhere in the middle of all of this, I realized the experience was a pretty good metaphor for artificial intelligence.

The technology can be incredible, but the figurative last mile, or two, can still be surprisingly difficult.

This week at 40 Accounting, we held an internal AI roundtable where our team shared how we are using ChatGPT, Claude, Gemini, Plaud, and other tools. We talked about what is working, what is not working, what saves time, and what occasionally creates more work than it eliminates.

What impressed me most was not the technology itself.

It was the learning.

Our team was sharing workflows, asking better questions, teaching one another, and figuring out how to turn incredibly powerful tools into practical improvements.

That is where the real value of AI begins.

It is not enough to have access to great technology. The real question is whether you can integrate that technology into your business in a way that makes your people faster, smarter, and more effective.

Ask yourself:

• Can we take a task that used to require an hour and complete it in 10 minutes?

• Can we improve quality while reducing friction?

• Can we make the process repeatable instead of reinventing it every time?

That is the last mile. Or, in our case, the last two miles.

I believe the companies that create the most value from AI will not necessarily be the ones using the most tools. They will be the ones that become exceptionally good at turning those tools into useful, repeatable workflows.

So here is your Strategy 4 Saturday take away: Do not simply ask whether your company is using AI. Ask whether AI is actually making your company better.

Pick one process. Find the friction. Test something. Measure the result. Share what works. Then do it again.

Technology is incredibly powerful, but execution still matters.

Because no matter how advanced the vehicle is, you still need to make sure it is taking you in the right direction, in the most efficient way possible.

Preferably without charging you $45 to go two miles through a maze of San Francisco side streets.

George Sandmann is a 30-plus-year veteran entrepreneur, author, and inventor, and the founder and CEO of Growth-Drive. He built his career on a question most owners get wrong: what actually makes a business valuable?

His answer is a framework he created called Strategic Capacity — a score for a business's ability to generate predictable profits, consistent cash flow, sustainable growth, and transferable value. It's gaining traction as a standard among business growth advisors, exit planners, CPAs, fractional CFOs, M&A pros, and private equity groups — and it's one of the core tools 40 Strategy uses with its own clients.

George is the author of The Growth-Driving Advisor (Forbes Books, 2023), a playbook for leading businesses from stuck to best-in-class. His central message cuts against the grain: revenue alone doesn't create value. A business that depends on its owner isn't yet an asset. The real work is building a company that runs,  and grows in worth,  without you.