Agentic Commerce: What It Means for Your Retail Media Campaigns

Something has quietly changed in how consumers make purchasing decisions - and most retail media campaigns have not caught up.

A shopper no longer types "French wine under £20" into a search bar. They open ChatGPT and ask: "I am cooking duck confit for a dinner party, my guests prefer lighter reds - what should I serve?" The AI recommends a wine. The shopper buys it. No sponsored search result was seen. No banner was served. No impression was recorded on any retail media network.

That is agentic commerce. And it is already happening at scale.

The Signal Problem

If you are running retail media campaigns across Amazon, Tesco, Boots, or any other network, your performance data starts at the moment a shopper arrives on the retailer's site.

What you cannot see is everything that happened before that - the research, the comparison, the recommendation that shaped the decision. That upper-funnel activity has always been partially invisible. But it used to live on Google, on category pages, on comparison sites - places where advertising could still reach it. Now, a growing share of it lives inside AI conversations. And that is a blind spot that no single retailer's dashboard will show you.

Adobe Analytics tracked a 4,700% year-over-year increase in AI-driven traffic to US retail sites in 2025.^[1] The volume is there. The attribution is not.

Three Things Disappearing From Your Data

Search queries. Shoppers are asking AI agents nuanced, contextual questions instead of typing keywords into retailer search bars. That intent data never reaches the retailer - and never reaches you.

Browse behaviour. The consideration phase - time spent comparing products, reading reviews, weighing options - now increasingly happens inside the AI conversation. Your category page analytics are showing you less of the story than they used to.

Ad impressions. If an AI agent is doing the browsing on a shopper's behalf, the human never sees the sponsored placements that would typically influence their decision. This is what the industry is starting to call Dark Search - the retail media equivalent of dark social. The purchase happens. The path to it is invisible.

What This Means for How You Buy Media

The instinct is to treat this as a future problem. It is not. The shift is happening now, and it has two immediate implications for how you manage retail media.

First, your upper-funnel metrics are becoming less reliable. If you are using search impression share, browse rate, or category page traffic as proxies for brand consideration, those numbers are starting to undercount. Brands that look flat in the data may be winning in AI recommendations - and brands that look strong may be invisible where decisions are actually forming.

Second, fragmented campaign management makes this worse. If your retail media activity is spread across five networks with five different dashboards, you have no way to see the consolidated picture clearly enough to spot where AI-driven dark search is eroding your signals. The gaps only become visible when the data is in one place.

The New Discipline: GEO Alongside SEO

There is a practical response to this beyond just better measurement. It is called Generative Engine Optimisation - GEO - and it is the discipline of ensuring your brand is the answer AI agents give when consumers ask questions in your category.

Where SEO was about ranking on Google, GEO is about being cited by ChatGPT, Perplexity, and Gemini. The inputs are different: structured product data, authoritative third-party mentions, complete and consistent SKU attributes, expert reviews. If an AI agent cannot parse your product clearly, it will not recommend it. If your brand is not present in the sources AI draws on, you will not appear in the conversation.

GEO does not replace your retail media investment. It protects the upper funnel that feeds it.

What to Do Now

You do not need to rebuild your entire strategy. But there are three things worth doing today.

Audit your product data. Incomplete attributes, missing GTINs, inconsistent descriptions - these are not just housekeeping issues. They are the reason AI agents skip your products. Clean, structured, machine-readable data is the new creative asset.

Consolidate your reporting. You cannot manage what you cannot see. If your campaign data across retailers is fragmented, you will not be able to identify where AI-driven dark search is affecting your performance - or where it is working in your favour.

Start tracking AI referral traffic separately. It is small today. It will not stay small. Segmenting ChatGPT, Perplexity, and other AI referral sources in your analytics now means you will have a baseline when the volume becomes significant.

Agentic commerce is not a threat to retail media. It is a shift in where retail media needs to operate. The brands that adapt their data, their content, and their measurement will have a significant advantage over those still optimising for a funnel that is quietly changing shape.

Triffecta gives retail media buyers a single platform to execute campaigns across multiple networks, with consolidated reporting that makes the full picture visible. Are you thinking about what agentic commerce means for your campaigns?