The funnel is not broken. It has simply been handed to a machine.
For the past decade, digital marketers have been optimizing the same basic sequence: attract a shopper via search or social, guide them through a product page, reduce friction at checkout. That model assumed one thing: that a human being was doing the browsing. In 2026, that assumption no longer holds.
Agentic AI, systems capable of autonomously perceiving context, planning multi-step tasks, and executing decisions without waiting for human input, is transforming e-commerce from a browsing experience into a delegation experience. Instead of searching for products, shoppers are increasingly handing their intent to AI agents. The agent searches, compares, negotiates, and buys. The human approves or simply waits for the package.
This is not a feature update. It is a structural change to how commerce works.
From Search to Delegation
To understand what is shifting, it helps to look at the actual numbers.
In 2025, traffic to US retail sites from generative AI browsers surged 4,700% year-over-year. These users also spent 32% more time on-site and bounced 27% less than traditional visitors. AI chatbot traffic to US retail sites increased 670% year-over-year during the holiday season. The quality of AI-referred traffic is, in other words, categorically better than what search engines have historically delivered: shoppers arriving from AI services are 38% more likely to buy than those from traditional channels.
The scale of the opportunity is significant. US advertisers will spend $71.98 billion on retail media in 2026, up 18.7% from 2025. Within that, AI platforms are expected to account for $20.9 billion in retail spending in 2026, nearly quadrupling 2025 figures. Looking further ahead, McKinsey projects the global agentic commerce opportunity at $3 trillion to $5 trillion by 2030, with up to $1 trillion in US B2C retail alone.
What Agentic Commerce Actually Means
The terminology matters here. Agentic AI is not a chatbot. It is not a product recommendation widget. Unlike chatbots that respond to individual queries, agentic systems combine memory, tool access, and multi-step reasoning to handle transactions end-to-end.
AI is evolving from a passive tool that offers prediction, to an active, autonomous resource that can execute complex, multi-step, prescriptive actions across every consumer and operational touchpoint. The practical implication: a shopper no longer needs to visit your store. They need only instruct their agent. The agent does the rest.
Major platforms have already moved decisively in this direction. Amazon, Google, OpenAI, and Meta have all launched AI shopping tools, while retailers like Shopify and Walmart navigate how much access to grant external agents. Google has created a shopping agent that can generate a shopping list from a handwritten recipe and automatically purchase the items. Amazon offers sponsored prompts inside its Rufus shopping assistant. Walmart is embedding advertising into its Sparky chatbot.
The conversation itself has become the funnel.
The Death of the Traditional Retail Media Unit
For brands and agencies working in retail media, the implications are uncomfortable.
AI shopping agents that bypass traditional search and browse behavior reduce the value of sponsored product placements, display ads, and keyword advertising. When an agent makes a purchase decision without the shopper ever seeing a banner or a sponsored listing, the traditional retail media unit becomes irrelevant.
This is not a future problem. It is a present one.
55% of US advertisers already report inconsistent targeting and attribution from retail media networks. Attribution was already strained. In an agentic world, where the decision is made inside a language model’s reasoning process rather than on a product page, traditional click-based attribution collapses entirely.
The new challenge is visibility at the point of inference, not the point of click.
GEO: The New SEO
The strategic response that has emerged is Generative Engine Optimization, or GEO. GEO is the practice of structuring product data and brand content so AI agents can discover, understand, and recommend it. As agentic commerce shifts product discovery from search engines to AI platforms, GEO has become a strategic priority.
The logic parallels the early days of SEO, with one critical difference. Shopping agents don’t return ten blue links and let shoppers choose. Instead, they select a single winner based on technical compatibility. First-mover advantage here isn’t incremental, but structural.
What does this mean in practice? Retailers must optimize product data architecture to ensure consistency, structured taxonomy, and agent-readability across all digital touchpoints. They must establish real-time data synchronization for inventory availability, dynamic pricing signals, and promotional mechanics.
Put plainly: if your product catalog is not machine-readable, clean, and structured, AI agents will not recommend you. You will not appear in the consideration set at all. Your backend has become your storefront.
The Retail Media Opportunity Within Agents
It would be a mistake to read this shift as purely threatening to retail media. The opportunity runs in both directions.
Agentic interfaces present retailers with a new environment for advertising and affiliate marketing. Behind the product recommendation algorithm, each retailer can decide on the rules and potentially offer sponsored placements, or a ranking influenced by a cost-per-click model or a commercial agreement.
These agent-powered recommendations can be significantly more personalized, as retailers leverage their unique advantage: unified first-party data signals from both in-store and on-site purchase history, along with membership information. By reconciling shopper identities across these channels and interpreting behavioral patterns in real time, retailers can deliver hyper-targeted product suggestions that reflect the complete customer journey.
The brands that will win are those who structure their data now to be discoverable within these systems, and who build relationships with the retail platforms developing agent-native ad formats.
Consumer Readiness: The Missing Piece
Technology readiness is one thing. Consumer readiness is another. The data here is nuanced.
73% of consumers are already using AI in their shopping journey, embracing AI assistants for product ideas, review summarization, and price comparison. While only 13% say they have completed a purchase after being referred by an AI assistant, 70% are at least somewhat comfortable with an AI agent making purchases on their behalf.
There is a gap between comfort and trust, and that gap is where brand strategy still has significant leverage. Shoppers are willing to delegate purchasing to agents, but they want transparency, auditability, and control. Brands that build AI-agent-friendly infrastructure while maintaining visible brand identity will be better positioned than those who optimize purely for agent discovery.
73% of consumers expect brands to use AI to better understand their needs. The expectation of AI-mediated personalization is already mainstream. The question is whether brands are ready to meet it.
What This Means for E-commerce Marketers
The transition to agentic commerce is not happening uniformly. While 88% of organizations now report using AI in at least one business function, most are still in the experimentation or pilot phase, with only about one-third scaling AI programs across the enterprise. 62% of respondents are experimenting with AI agents, and just 23% have begun scaling agentic AI in any function.
For e-commerce marketing teams, the practical priorities are clear:
First, audit your product data. Every product in your catalog needs clean, structured, machine-readable attributes. Missing data, inconsistent taxonomy, and outdated inventory signals are not just SEO problems. They are agent-exclusion problems.
Second, think in terms of infrastructure, not campaigns. The marketing investments that will compound over the next three years are not in ad formats. They are in real-time inventory APIs, structured data feeds, and compliance with emerging protocols like Google’s Universal Commerce Protocol.
Third, reframe the retail media conversation with clients. The question is no longer how to win a sponsored placement on a search results page. It is how to ensure your brand is in the recommendation set when an AI agent acts on behalf of a high-intent shopper.
The agents are not a new channel. They are a new consumer surface, on par with what Google Search was to discovery twenty years ago. The difference is that the conversation itself is the funnel. 2025’s agents focused on product discovery and purchase. 2026’s agents will handle the full customer lifecycle.
The brands that understand this now will not simply capture more AI-referred traffic. They will define category authority for the next decade.