AI Native on the Earnings Call

Walmart’s US CEO used the phrase “AI native” on a public earnings call this week. Not in a press release or a keynote — on a call where analysts move billions based on word choice. He didn’t hedge it.

Then came the numbers. Their shopping agent, Sparky, doubled weekly active users in a single quarter. Average order value for Sparky users runs 35 percent higher. Units moved through the agent grew 4x.

I wrote about this tonight because the pattern is the one I keep seeing in my own work. Ship before it’s perfect, measure in real units, let the system compound. Sparky’s response quality improved 40 percent this year — meaning it was significantly worse when they launched it. They shipped anyway. The data made it better. Waiting would not have.

What stuck with me isn’t the scale. It’s the switching cost. A customer whose meal plans, purchase history, and preferences live inside an agent isn’t comparing apps anymore. That’s a moat built on workflow, not model quality.

Most companies are still debating which tool to buy. Walmart just reported Q1 results from a system that’s already compounding. The distance between running AI and talking about AI got measurable this week. Publicly, on the record, in front of Wall Street.