“This year it goes from answering questions for you to doing things for you in the real world.” — Kevin Weil, OpenAI Chief Product Officer, March 2025

OK, this quote is more than a year old, which is roughly an eon in AI time. And Weil is no longer even at OpenAI, but the point still holds: For most of us, keeping up with the pace of AI change is like trying to keep up with a world-class marathoner. Pointless.

For the past couple years, most of us have thought about artificial intelligence as something we chat with. Ask a question. Get a blindingly fast answer. Sometimes it helps. Sometimes it’s verbose and bland. We’ve learned to ask better questions, provide better context and, occasionally, get an answer that makes us think, “OK, the world really is changing fast.”

But the bigger and more recent shift goes far beyond AI that answers questions. It is AI-assisted systems that help complete work.

One of the great truths of run specialty retail (and any small business) is that there is always more work than time. Every day forces a tradeoff. If we do one thing, we are not doing something else.

AI-assisted systems give us a way to reduce the repetitive parts of the work and redirect that time toward higher-value action. Less generating, compiling, formatting and interpreting reports. More time thinking and acting on the data. More time with the team. More time in the community. More time for running?

Earlier this year, I wrote about the prospect of flipping the scarcity of time to an abundance of time. Maybe the better framing is this: AI assisted systems give us a chance to repurpose time toward the work that matters most.

No robots, more data

The immediate future of AI in running specialty probably won’t involve a robot shopping the internet for us. It is more likely to begin with something much quieter, like a nightly ledger that turns an unruly pile of daily retail data into decision-ready information.

One approach we have taken is to use an AI-assisted system to compile a daily footwear activity ledger. While not truly “agentic,” it is still automated and run by software that was built without any coding skills.

With the ledger in place, a lot becomes possible.

Rather than running reports, we can now “talk” with the data just by asking questions directly to it. An even more useful trick is quickly turning the data into dashboards for weekly and even daily (yes, daily!) at once ordering. This same data becomes monthly manufacturer summaries, KPIs and even special sourcing trend sheets exposing inventory gaps to address.

The data is no longer trapped in the POS waiting for someone to carve out time to build the report. It’s delivered to us, automatically (via email), in a format and at a cadence that allows for fast and impactful decisions.

(As an aside, I sometimes feel like George Jetson when I catch myself complaining that I had to press too many buttons to get to the data. For those who know this reference, I see you smiling).

And the most recent breakthrough is using this data to help write footwear futures orders.

Write less, think more

As I write this, we are deep in SS27 order-writing season. Anyone who writes footwear futures knows the time-consuming work starts long before the truly important work of building the order. There are reports to be run. Data must be compiled. Spreadsheets need to be formatted. Prior seasons compared. Order sheets to be built.

By the time everything is ready, the buyer may have little energy left for the part that matters most: thinking.

Historically, I might spend 70 percent of my time preparing information and building orders, with the remaining 30 percent actually thinking about the order. What we are building has the potential to flip that ratio. Now, the process may be 70 percent thinking and 30 percent preparation. Maybe even 80/20.

We do not lack decisions to make. We lack the time and clean information needed to make them well. AI-assisted workflows can change this by making more room for judgement. They remove some of the repetitive preparation that keeps us from using our judgment in the first place. If we think AI only means a chatbot answering product questions, we miss a much bigger opportunity.

Start small, grow with it

The first step doesn’t have to be grand. Pick a recurring task that consumes too much time but improves the business when done well. Automate the gathering of data. Standardize the format. Automate the delivery of the result. Then use your expertise and taste to make better decisions. If you don’t know how to do this (most of us don’t) just ask your preferred AI model (ChatGPT, Claude or Gemini). With some patience and persistence, it can guide you through the process.

The goal is not to remove humans from the work. The goal is to enable humans to spend their precious time doing the work that matters most. We’ve entered a new era in the fast-paced AI evolution. It’s time to build AI assisted systems and workflows that make us better at what we do best.

If you are experimenting with similar workflows, I’d love to compare notes. I’ll be at The Running Event in San Antonio, TX, in December and would welcome the chance to connect there or in advance. You can reach me at [email protected].