What a Marathon Taught Me About AI and What AI Could Never Replace

Many miles. Many lessons. And a reason to finish that had nothing to do with my time.

My dad died of cancer eighteen months ago.

 

When I registered for Grandma's Marathon, I told myself it was about unfinished business. Twenty-three years earlier I ran my first marathon, got injured during the race, hobbled across the finish line, and missed my goal by fifteen minutes. That never fully left me.

 

But as I went on the journey of training, collecting donations, and getting to the starting line, something shifted. This wasn't just about settling an old score. It was about honoring him.

 

I ran with Team Determination for the American Cancer Society. Together our team raised more than $130,000 for the Minneapolis Hope Lodge, a home away from home for people facing cancer.

 

That's the reason I was there. Everything else is just the story of how I got to the starting line.

When I trained for my first marathon, I followed a plan torn from a running magazine. I used my car's clock to time my runs. If I wanted to know how far I had gone, I either guessed or drove the route to measure it. I wasn't particularly disciplined, but what I lacked in discipline I made up for in youth and energy.

This time I had a smartwatch tracking every mile, heartbeat, and metric. I also had AI.

Initially, AI did exactly what I expected. It helped me build a training plan that accounted for my schedule, fitness level, age, and goals. I work with AI every day, so this was useful, but not surprising.

Then I got hurt and AI became something different.

The injury didn't change the goal. It raised the stakes.

I couldn't honor my dad as planned if I dropped out and I wanted to keep the commitment I had made to everyone who donated.  Before I could think about pace or time, I had to solve a simpler and harder problem of getting to the starting line healthy.

My partnership with AI evolved into something far beyond a souped-up search engine. It became a data warehouse, a strategy partner, a coach designed to keep me from making the injury worse. Every morning started with a report and a conversation — how the Achilles felt, what had changed, what the data suggested. I would push back, add context, and we would align on a plan. Then execute.

The conversations were always anchored to the same goal: get to the starting line and finish the race.

That alignment kept my training on track and kept my ego in check. It also insulated my wife, Jenny, from a few hundred more miles of marathon talk.

Those long training miles gave me a lot of time to think. About AI. About the conversations I was having with business leaders while I was using it for something personal. About how similar the challenges were.

Here is what I kept coming back to: the ceiling on what AI delivers is set by the person using it, not by the technology itself. Goal clarity, data quality, critical judgment — every one of those variables is human. The technology performs to the level of the operator.

The same lessons kept surfacing in both worlds.

The fastest way to make a problem worse is to panic and push harder.

The advice I least wanted to hear was the advice AI kept delivering: stop trying to make up the miles you missed, don't push your pace. It didn't get impatient. It didn't get emotional. It stayed locked on the goal when I wanted to override it with stubbornness and optimism.

Ambitious people always want to push harder. But the behavior that creates the setback is often the same behavior that prevents the recovery.

I've watched companies do the exact same thing. When goals aren't being met, the instinct is to do more, faster. That's exactly how a manageable problem becomes an expensive one. AI will confidently give you more options faster.  That doesn't mean you should use them. The discipline to stay anchored to the goal, make informed decisions, and focus energy on what will create the biggest impact is as hard in business as it is at mile eighteen.

Good data gets you to the starting line. People get you across the finish line.

AI amazes me with its insights and power. It also frustrates me with its flaws. It would help me analyze training data precisely one session and confidently reference today as yesterday in the next.

People talk a lot about the importance of data input quality — and they're right. But that's only half the job. Someone still has to critically evaluate what comes out. It's a yes and: garbage in produces garbage out, but clean data going in doesn't guarantee good decisions coming out.

I recently sat with a C-suite exec at a forward-thinking tech company who told me he feared his team was submitting AI-generated reports and presentations without really reading them. The tool had become a shortcut rather than a partner. When professionals stop engaging critically with AI output, they haven't created efficiency. They've transferred accountability to a system that sometimes doesn’t know what day of the week it is.

Data quality gets you to the starting line. Judgment gets you across the finish line.

AI impact is a journey. The further you go, the greater the advantage.

Most people start by using AI as a better search engine. That's where my training work started too. Ask a question. Get an answer.

But that's not where the value ended up.

As the marathon became more challenging, my use of AI evolved. It stopped being a tool for finding information and became a tool for making decisions. It understood my goal, training history, injury, and the tradeoffs I was navigating.

That evolution is happening inside companies right now.

The first wave of AI adoption is about individual productivity. The second is about organizational automation. Both create value, but the organizations pulling away from the competition are moving beyond answers and tasks.  They’re using AI to build context, organizational memory, and ultimately better decisions. That's the progression behind the AI Operating Model I've written about previously (Link here)

The technology becomes more valuable as its understanding deepens. But the ceiling never changes. The quality of the outcome is still determined by the capabilities of the humans using it.

Tim and Andrew were my pressure test. Without them challenging what AI was telling me, I would have followed confident recommendations down the wrong path more than once. The tools around you — human and artificial — only work as well as the judgment you apply to them.

The turning point came after my longest training run. Nineteen miles.

I was exhausted and sore and, for the first time in weeks, not worried. The injury hadn't returned. The data and my body were finally telling the same story.

I wasn't where I wanted to be. But I was healthy enough to race.

Race day in Duluth started gray and cool, then opened into a beautiful North Shore morning — the kind that makes you remember why people run this course.

By mile 14, I knew the clock wasn't going to say what I had originally aimed for. The missed weeks, the cross-training that kept me fit but couldn't fully replicate the miles, the Achilles that had held but taken something with it — it all showed up eventually.

What I didn't expect was the weight of the final miles.

I thought about my dad. A lot. In the stretches where the race got hard and the finish line felt impossibly far, he was what kept me moving. Not the time on the clock. Not the training data. Him. The feeling that crossing the line meant something that no finishing time could measure.

Crossing the line, my first thought was honest: thank God it's almost over.

My second thought was about him.

Here is what two marathons, twenty-three years apart, taught me about AI.

The technology is genuinely powerful. It created real efficiency. It helped me train smarter, adjust when conditions changed, and stay focused on what mattered most when ego was pulling me toward bad decisions.

But AI didn't get me to the finish line.

Tim and Andrew did with knowledge and experience that no algorithm could replicate. Jenny did with patience, presence, and the steady support. My dad did by giving me a purpose that went far beyond miles and minutes.

AI amplifies what you bring to it. The clearer your goal, the more honest and consistent your inputs, the better it performs. Using it well is a skill.  But it has no access to what actually drives you. It cannot tell you what matters. It cannot give you a reason to keep going when the race gets hard.

That part has always been human. 

I crossed the finish line two days before my 50th birthday, in honor of my dad, and helped raise meaningful money for a cause that's important to me.

No clock could measure that.

I write about AI strategy and go-to-market at www.sandgtm.com — including the AI Operating Model referenced above.