What keeps commerce businesses valuable as customers shop through AI?

AI services can gather a buyer’s needs, compare offers and become the place where a purchase begins. Commerce platforms, retailers and service providers can continue to supply the systems, goods and operations behind it. A change at the interface can therefore redistribute influence without transferring every operating role.

This opens several positions at once: organising demand, connecting sellers to channels, executing transactions, and providing something buyers specifically want. Firms can hold more than one. Payment, identity and customer data cut across them; they do not move automatically with the purchase screen.^[7]

The result is a contest over how those contributions fit together. An AI service may shape which provider gets considered. A platform may connect merchants to several services. A provider may keep its own account, basket or checkout inside an AI experience. These arrangements create different opportunities to influence choice, maintain a relationship or become replaceable.^[6]

The important change is therefore broader than “shopping in a chatbot”. Activities can be separated and recombined across firms. That makes the boundaries between businesses—and the capabilities that sustain those boundaries—open to reconsideration.

Not every partnership moves the customer route. Retailers such as Lowe’s run their own assistants on lab models: they keep the interface and assortment and import capability, taking on a dependency on the model supplier instead of a channel. The same lab can play both roles—Google supplies Home Depot’s and Ulta’s assistants and runs a channel where their products compete.^[1]

A simple item purchase, a grocery basket and a delivered meal require different coordination. A channel that can capture an order still needs someone to accept it, keep it feasible and deliver the service. Which firms remain involved depends on the activity system—not just the sophistication of the model.^[3]

The restaurant research illustrates this distinction. Pickup can connect an AI channel to a point-of-sale platform without the same courier requirement as delivery. That creates a possible route around a delivery marketplace for one activity while leaving a need for delivery capability elsewhere. A marketplace is exposed when customers use that direct route for orders it would otherwise coordinate.

Provider differentiation also matters. Buyers who ask for a particular brand, account benefit or service give that provider a different position from a seller presented as one interchangeable option among many. Retaining an account can help sustain that distinction; it does not, by itself, establish who wins the next purchase.

Standards add another tension. A common interface can expand a platform’s reach while making its connecting work easier for others to reproduce. The platform’s position then depends on what it supplies beyond translation: execution, a merchant base, data quality or coordination across channels that still differ.^[4]

For a commerce platform, the question is whether it remains valuable as more connections become standard. Expanding into more AI channels can grow its role, but the same openness can support rival connectors or direct participation. Its position is stronger when it coordinates a service that merchants and AI channels would find costly to reproduce through another connector.

For a retailer or service provider, the question is what gives buyers a reason to choose it even when discovery happens elsewhere. That may be a distinctive product, an account, a basket, service quality or a trusted relationship. The strategic choice is what to make accessible through partners while continuing to develop that distinction.

For an AI service, shaping selection creates an opportunity. Turning that opportunity into durable leverage would also require buyers to return, providers to be substitutable and alternative access to be weak. ^[5]

No single outcome follows. Several arrangements can coexist, and AI can remain an additional channel with limited redistribution of advantage. Strategy should distinguish commitments that help across several futures—strong capabilities and viable alternatives—from bets that depend on one route becoming dominant.