Companies are racing to build AI-shopping tools now, even though those agents barely move money today — why the rush?

This explores why companies are rushing to build for AI shopping agents even though the money those agents move today is small. The corpus doesn't measure today's agentic revenue directly. What it does show is that the urgency is mostly about position: who sets the rules, who holds the customer relationship, and who gets locked out. Those things get decided early, long before the revenue arrives. The headline numbers are about 2030, not now. McKinsey's figure of up to $1 trillion in US retail is a 2030 projection Could agentic commerce reach one trillion dollars by 2030?. BCG estimates that $1.3 trillion of e-commerce spending will shift to agent-assisted or autonomous shopping, with consumers and platform networks capturing most of the gains and merchants at a structural disadvantage Who wins and loses as AI agents reshape shopping?. If you expect to be on the losing side of that shift, waiting for revenue to justify your investment means arriving after the terms are already set.

One formal model makes a striking point: platforms and consumers judge agentic search by different measures. Platforms rank it by conversion revenue alone, while consumers also pay the cost of searching. As a result, platforms rationally want agentic search before consumers rationally should switch to it Why do consumers delay adopting agentic search?. So the gap in the question isn't confusion. Supply is built to run ahead of demand. Early signals suggest demand is starting to catch up in small but real ways. On the x402 payment rails, the median time from funding an agent wallet to its first purchase dropped from 73 days to 18 minutes, and repeat funding rose to 89% Are AI agent wallets becoming genuinely operational instead of experimental?. The volumes are tiny, but the behavior is turning into habit, and habits are what incumbents fear missing.

The bigger driver is standards and lock-in. Agentic commerce protocols (ACP, AP2, UCP, Visa's TAP, Web Bot Auth) work as stacked layers rather than competing alternatives Do agentic commerce protocols compete or complement each other?. Waller expects the market to split into open and closed systems, much like open- and closed-loop payment networks, and that split will decide who controls the market Will agentic commerce platforms adopt open or closed models?. The UK's CMA argues that without data portability, consumers get locked into incumbent ecosystems and agents can't act effectively on their behalf Should data portability be infrastructure for agent-driven markets?. Standards and data access harden early. Firms that join later get the terms others wrote.

The less obvious stake is preference itself. Bilgihan and colleagues argue that agents that complete purchases, not just recommend them, help shape what consumers want through repeated conversations Do AI shopping agents reshape what consumers actually want?. Whoever owns the agent partly shapes demand. For some industries the urgency is defensive. McKinsey argues that agents moving cash and choosing card rewards could wipe out the profits banks make from customers who never switch, tightening deposit margins by 30–50 basis points if just 10–20% of consumers adopt cash-sweeping agents Will AI agents break the deposit and credit card inertia?. For a bank, the urgent number is the revenue it could lose, not the revenue agents might bring in.

So adoption urgency and near-term opportunity measure different things. One tracks control over standards, data, customer preferences and existing profits. The other tracks this year's transactions. Building early also has a practical side: the IMF argues that payment systems need to keep the AI's probabilistic decisions separate from the predictable, legally enforceable authorization and settlement steps Should AI agents make payment decisions independently?. That plumbing takes years to build and has to be ready before trusted volume can arrive.