Retail promised fast, on-demand delivery before it had a way to pull it off. Food apps spent a decade training all of us to expect it, and that expectation has spread to paint and lumber, bulk groceries, and things nobody ever planned to move in under an hour.

This past year, Sam’s Club switched on one-hour delivery across 600 clubs and pushed 65,000 of those orders out in a few weeks. Uber Eats started sending drivers to collect retail returns off people’s porches. Yet while those numbers will likely be stale by next quarter, the shift they point to won’t be. In areas that were never built for it, fast delivery has gone from nice-to-have to expected.

Underneath it is software making thousands of small calls an hour, deciding who picks up each order and how it reaches the door. The fix everyone reaches for is the AI agent. Most of what’s sold today isn’t ready. As of June 2025, Gartner counts only about 130 legitimate vendors among the thousands now selling agentic AI, and expects more than 40% of agentic AI projects to be scrapped by the end of 2027.

The whole argument fixates on the wrong thing. Everyone wants to know what the AI can do. But in spending years connecting retailers to delivery networks, I’ve learned to ask something narrower: What has this software earned the right to decide in your operation?

One trap is what Gartner calls agent washing : a chatbot or rules engine relabeled an “AI agent” and priced like one.

Source link