Amazon, Google, Walmart, OpenAI, Shopify, and other global players are no longer just competing for ad spend or traffic. They are racing to control the AI-powered shopping agents that will make purchasing decisions for millions of shoppers. Beyond capturing search or checkout, the winner will own the entire shopper journey, from discovery to loyalty, without the shopper ever touching a web page.
Answer Engines Are the Leading Indicator
AI-powered answer engines that can search, compare, and act on behalf of the shopper are already reshaping retail.
- ChatGPT now serves nearly 1 billion weekly active users, up from zero less than two years ago. Commerce queries are among its fastest-growing use cases.
- On July 8, the first day of Amazon’s summer Prime Day event, this year, generative AI queries to US retail sites surged 4,100% as shoppers asked AI to find deals, compare products, and check stock in real time.
- Google’s AI Overviews have already surpassed 2 billion queries, with retail-related searches emerging as a dominant category.
From “First Click” to “First Conversation”
Ecommerce has always been about being the first site visited and the last site visited before checkout. That advantage is disappearing fast. In the agentic commerce era, shoppers will simply tell their AI what they want, and the AI will handle discovery, comparison, and purchase without the shopper ever visiting a retailer’s site or app.
The new “front door” to commerce is a trusted AI agent embedded in the shopper’s phone, browser, operating system, or favorite retailer’s app, not a homepage or product listing page.
The Strategic Stakes
Every player sees the same prize: a personal AI that knows a shopper’s style, budget, dietary needs, and brand loyalties, and buys instantly on her behalf. But delivering that vision requires deep access to product catalogs, inventory, and pricing. Giving that access to an outside platform means ceding customer relationships, a risk most retailers are unwilling to take.
That is why the AI commerce wars are as much about defensive blocks as they are about bold launches:
- Amazon cut off Google’s Mariner shopping bot and pulled shopping ads from Google, signaling that it will not feed a rival’s AI. Meanwhile, Amazon’s Rufus conversational shopping assistant and Buy for Me enable cross-retailer purchasing within Amazon’s ecosystem.
- Shopify is developing a universal cart for millions of merchants, keeping a human in the loop for brand control.
- Walmart is rolling out multi agent ecosystems spanning shopper, associate, partner, and developer agents.
- OpenAI is testing Operator, an AI agent that can buy across the open web, while integrating shopping into ChatGPT.
- Google is launching AI Mode in search tied to its Shopping Graph.
- Apple is embedding AI agents deep into iOS and macOS. Microsoft is doing the same with Windows and Edge.
Retailers are moving beyond old defenses like robots.txt to server-side controls, content delivery network-level blocking, and API-gated access. The next frontier is permissioned data sharing, deciding which AI agents are “friendly” and feeding them curated, enriched product data that works in the retailer’s favor.
The likely endgame is walled gardens populated by “super agents” such that:
- Amazon’s Rufus will interpret only Amazon’s catalog.
- Google’s Gemini agents will live off its Shopping Graph.
- Walmart’s agents will operate inside Walmart’s closed ecosystem.
- Shopify will offer a merchant-friendly environment, but on its terms.
The winner won’t just have the smartest AI. It will have the richest, most exclusive view of the shopper.
Implications for FMCG Brands
- The traditional marketing funnel is irrelevant in a world where agents compress it into a single millisecond.
- AI agents are a new distribution channel.
- CPG advertising becomes less about storytelling and more about data signaling.
- Trade marketing must adapt to agent-first decisions, where structured product data drives inclusion in an AI shortlist.
- Attribution will get harder as a single agent consults multiple data sources before purchase.
- Brands must operationalize loyalty across two strategic fronts: one emotional, one algorithmic.
- Differentiation will depend on metadata marketing, codifying brand value in structured data like ESG claims, verified reviews, and precise product attributes.
- Brands that build agent-friendly infrastructure now can secure an outsized share of voice and revenue on the emerging “algorithmic shelf.”
- Retail media network reach may shrink if AI agents bypass retailer sites.
What FMCGs Must Do Now
- Build an agent-ready product data strategy. Ensure every SKU has enriched, machine-readable attributes (pack size, nutrition, sustainability, pricing tiers).
- Shift to metadata marketing. FMCGs must structure every product detail so AI agents can instantly find, compare, and recommend, turning precision data into a new source of demand.
- Negotiate access early. Secure your place in retailer or marketplace AI ecosystems before exclusivity rules lock out competitors.
- Integrate with “friendly” agents. Partner with platforms that align with your retail and brand strategy.
- Rethink retail media KPIs. Go beyond clicks to measure agent-triggered adds-to-cart and influence on purchase decisions.
- Experiment with AI-detectable promotions. Test instant savings, multi-pack discounts, or bundles that agents can detect and act on in real time.
Agentic commerce will redefine the competitive map of retail. Successful platforms will combine AI intelligence with exclusive data access, brand trust, and operational integration across the shopper journey. The early winners will set the rules of engagement for brands, shaping how product discovery, pricing, and loyalty work in an AI-first world.
For FMCG brands, the next 12-18 months will be critical. Those that invest now in agent-ready data, controlled access partnerships, and AI-detectable value propositions will position themselves as preferred choices in the AI shortlist. Those that wait risk becoming invisible in a retail world where the conversation, not the click, drives the sale.