Over the past few weeks I’ve had several conversations about Agentic Commerce with several senior executives at large retailers and CPG brands and there is clearly a significant amount of interest. First let me clarify that by Agentic Commerce I (and most of them) are referring to the entire consumer journey from discovery to purchase and not just the transaction (which I refer to as Agentic Shopping).

The amount of interest in LLM based commerce has increased exponentially — as one brand put it — “we always go where the consumer is” and with ChatGPT reporting that their weekly active user count is running at about 800M, clearly LLMs are where they want to be.

The problem is — how do you get there? One of the points of confusion right now is that protocols and standards like ACP (OpenAI, Stripe) and UCP (Google, Shopify and others in a consortium) means agentic shopping. It doesn’t — the standards (along with MCP and AP2) are meant to be the “Rails” on which Agentic Commerce is being built. So whether a retailer or brand wants to keep control of the final transaction or wants to enable direct purchases in the LLMs, these are essential protocols for Agentic Commerce.

Adding to this confusion is GEO — GEO is simply put, a way brands and retailers can ensure better discoverability by making their content easily crawlable and enabling it to be indexed into “Answers” within the LLMs. This however while it is Agentic Commerce 101 — the table stakes and not betting on the full game, it is simply not sufficient to capitalize on the enormous amout of shopper traffic LLMs are gaining.

OpenAI, Gemini and Microsoft Copilot respectively have introduced Apps, Business Agents and Copilot agents which when implemented, work to increase visibility for brands inside the LLM. These technologies (which can be used to implement storefronts) are powerful ways for retailers and brands to enter the consumer conversation, offering up relevant content and answers. They are however just technologies and require diligent curation of content and guardrails to add significant value over generalized answers from the LLMs. Of the three, Open AI seems to be the most robust and optimized towards conversational commerce. This is the next frontier in Agentic Commerce.

Key takeaways from my conversations are:

  • Retailers are not (yet) rushing into Agentic Shopping but seem very interested in increasing discoverability and being part of conversations in LLMs but ideally want consumers to buy from their site/apps and are implementing chat within their sites/apps
  • Brands are investing heavily in GEO but want to go much further and see a real opportunity to “win the consumer, not just the transaction” and are thinking about ways to build a presence and be in consumer conversations
  • Both see Agentic Commerce as a critical investment to engage consumers on these new platforms. They are also eyeing LLM ads as a way to get there.