AI is turning e-commerce into a coordination problem

E-commerce used to be built around a fairly tidy idea: a shopper arrived, compared a few options, and checked out. Clean. Efficient. Almost suspiciously so. That model is now looking a bit dated.

Shopping is increasingly spread across assistants, feeds, chat surfaces, and live video. Each surface seems to hold a different piece of the decision. One channel helps with discovery, another with evaluation, another with recommendation, and another with the final nudge. The buyer is still buying the same item, but the path is no longer a straight line.

From funnel to relay race

The merchant’s job is changing with it. Instead of managing a single funnel, they are managing something closer to a relay race, where the baton is product identity. The same product has to be discoverable in ChatGPT, legible in Google’s shopping surfaces, recommendable in Meta’s AI and messaging tools, and still compelling inside TikTok’s content-led environment.

If the baton drops, the journey can fragment. Stale pricing, weak product data, inconsistent availability, or poor imagery may cause the shopper to re-enter elsewhere. In other words, the issue is not just whether a product is present. It is whether it stays coherent as it moves across surfaces.

“The merchant no longer controls a single funnel.”

What the platforms are doing

The shift is not about one channel replacing another. The evidence suggests something more layered: new entry points are being stacked on top of the same purchase process.

• OpenAI is pushing product evaluation deeper into the assistant layer.
• Google is making cart state persistent across surfaces.
• Meta is turning messaging and Reels into qualification and recommendation layers.
• TikTok is collapsing discovery and purchase into the same content stream.

Put together, that does not look like a single dominant channel. It looks more like a distributed commerce graph, with multiple routes feeding the same decision. The shopper may not notice the architecture. The merchant certainly will.

Why this changes the work

Channel strategy starts to look less like media buying and more like systems design. That is a useful shift, even if it sounds like the sort of phrase people say right before a meeting runs long.

The practical implication is straightforward: merchants need consistency across product data, pricing, creative, and re-engagement. These surfaces behave differently, but they feed the same decision. If one part is out of sync, the whole experience can feel disjointed.

This also means the quality of the underlying product information matters more, not less. AI-enabled surfaces may make products easier to surface, but they do not remove the need for clear, current, and usable inputs. A recommendation that feels generic or stale is still just a recommendation that feels generic or stale.

Not every merchant is starting from the same place

The benefits do not appear to be evenly distributed. Larger catalogs, richer data, and stronger operational discipline are likely to benefit first. Smaller sellers may find the coordination burden rising faster than the payoff.

That is an important caveat. The more surfaces that participate in the journey, the more work it takes to keep the experience aligned. AI may reduce friction in some places, but it can also raise the bar for operational discipline elsewhere.

So the headline is not that AI has replaced e-commerce. It is that e-commerce is becoming a coordination layer. The checkout page is still there, but it is no longer the whole story. The real work is happening earlier, across more surfaces, and with more moving parts than the old model ever had to manage.