Agentic commerce: what happens when AI starts shopping for you

For most of the internet's commercial history, AI was a behind-the-scenes player. It powered recommendations, sorted search results, and personalised email subject lines. The shopper still made the final call.

That is changing fast.

Agentic commerce is the shift from AI as a recommendation engine to AI as an active participant in the purchase process. McKinsey defines it as AI that "anticipates consumer needs, navigates shopping options, negotiates deals, and executes transactions." Not later, not in theory. Now.

ChatGPT launched Instant Checkout in September 2025, enabling purchases directly from chat with brands like Glossier and Spanx. OpenAI and Stripe built the Agentic Commerce Protocol (ACP) to make this infrastructure available across merchants. Google followed with its Universal Commerce Protocol (UCP) and Agent Payments Protocol (AP2). 900 million weekly ChatGPT users now have a checkout button available.

At Livewall, we design and build loyalty programmes, digital activations, and owned brand experiences for consumer brands. This shift makes all of that work more urgent, and more relevant.

How the mechanics work

When a consumer uses ChatGPT or a Gemini shopping assistant, they are not browsing a catalogue. They are having a conversation that ends in a transaction.

The AI acts as a personal shopper. It considers stated preferences but also inferred ones: past interactions, browsing signals, and behavioral patterns built up over time. The shortlist it shows you is already filtered through what it thinks you will actually want.

For merchants, the protocols work like this: brands that integrate with ACP or UCP become accessible to the AI layer. Brands that do not are effectively invisible to shoppers using AI agents for purchasing. This is not future risk. The infrastructure is live today.

The speed of adoption is striking. Consumer use of generative AI for shopping climbed from 38% in 2024 to 51% in 2025. Traffic from generative AI sources to US retail sites increased 4,700% year over year. These are not pilot programme numbers. The consideration phase of shopping is already moving into AI-mediated territory.

What it means for brands when AI is the intermediary

When AI acts as intermediary between a consumer and a brand, the traditional purchase funnel collapses. Discovery, consideration, and purchase happen in a single conversation. Product awareness, brand storytelling, retail theatre, all of those touchpoints are bypassed.

This creates two categories of brand: those the AI considers, and those it does not.

Getting into the first category requires more than good product data. It requires the kind of reputation signals AI systems weight: reviews, behavioral patterns, and evidence of consumer loyalty and preference. Brands that have built real relationships with their customers, through loyalty programmes, owned communities, and direct engagement mechanics, have something the AI can work with.

We have seen this with the brands we work with. HEMA Stapelgek created daily return habits through a gamified loyalty mechanic inside the HEMA app. Decathlon's always-on programme ties rewards to everyday physical behavior. These programmes generate behavioral signals that outlast any single campaign and create the kind of persistent brand preference that matters in an AI-mediated world.

What to build for an AI-mediated world

The brands that will navigate agentic commerce well share a few characteristics.

They have first-party data that reflects real customer preferences, not just transactions. They have built loyalty mechanics that connect with customers between purchases, not only at the point of sale. And they have created owned digital touchpoints that generate habitual direct engagement rather than occasional traffic spikes.

The shift does not make brand building redundant. It makes it more important. When an AI is deciding what to put in front of a consumer, the brands that win are the ones whose customers have expressed a genuine preference. Not to an algorithm. To themselves, through real interactions.

Gamified loyalty is one of the most effective ways to build that foundation. It creates interaction patterns that go beyond purchases, rewarding reviews, participation, community engagement, and advocacy. Each of these interactions builds the kind of behavioral profile that AI agents can use to surface your brand at the right moment.

At Livewall, we help consumer brands build these interactions. Programmes that earn real behavioral signals. Experiences that create genuine habit loops. The groundwork for a world where an AI recommending your brand is a reflection of something your customers genuinely chose.