This post was originally published on this site.
Google recently announced the Universal Commerce Protocol (UCP), a new open standard designed to accelerate the adoption of agentic commerce. UCP establishes a common language that allows AI agents to operate across the entire shopping journey—from discovery and conversion to post-purchase support. The protocol was developed in collaboration with industry leaders, including Shopify, Etsy, Wayfair, Target and Walmart, and is backed by more than 20 additional companies, such as Mastercard, Visa and The Home Depot.
The Collapse of the Traditional Shopping Funnel
Agentic commerce represents a transformation from traditional e-commerce. According to Adobe’s Digital Economy Index, traffic from AI sources to retail sites has increased 1,200%, underscoring a change in how consumers discover products. This evolution is driving a transition from search engine optimization (SEO) to answer engine optimization (AEO), where machine-readable product information and structured data outweigh traditional keyword strategies.
Gartner research suggests that by 2029, agentic AI will autonomously resolve 80% of common customer service issues, potentially reducing operational costs by 30% while significantly improving response times. This autonomy is also extending into the transaction phase. Google’s AI Mode, for example, now includes price-tracking and automated checkout capabilities. Once users save preferences such as size, color and budget, an AI agent can monitor listings, send price-drop alerts and complete purchases securely through Google Pay.
“Our biggest bet on this technology is agentic commerce,” said Juan Luis Bordes, VP & General Manager, PayPal, in an interview with MBN. “In this model, users will simply state their needs to an AI assistant, and the purchase will be executed seamlessly in the background.”
Taz Patel, Head of Advertising and Shopping, Perplexity, said that “some aspects of the AI future are already clear. Consumers want agentic experiences throughout their shopping journey, and they turn to Perplexity for accurate answers they can trust. When our systems can ingest clean, well-organized product information with rich attributes, consistent taxonomy, and up-to-date availability, the results speak for themselves: more relevant search experiences, higher conversion rates and better alignment with shopper intent.”
The Rise of Multi-Agent Systems (MAS)
A critical development in agentic commerce is the evolution from single AI tools to multi-agent systems (MAS). These collaborative networks of specialized agents work together to manage end-to-end retail functions. Gartner reports that 75% of organizations plan to deploy multi-agent frameworks within the next 18 months.
MAS applications span several core retail functions:
-
Dynamic Pricing: Agents analyze competitor pricing and inventory levels in real-time to optimize promotions.
-
Supply Chain Resilience: Autonomous agents coordinate with suppliers and reroute orders automatically during disruptions.
-
Inventory Management: Predictive agents forecast demand and trigger replenishment based on multiple variables.
Google Cloud’s Gemini Enterprise for Customer Experience (CX) integrates these capabilities into a single interface. The platform uses digital concierges with multi-step reasoning to understand customer intent across web, app, chat and phone channels. Companies including Kroger, Lowe’s and Papa John’s have begun deploying these tools to manage the customer lifecycle from discovery through post-sale support.
Walmart has also partnered with OpenAI to embed ChatGPT into its retail platforms, aiming to modernize the shopping experience and improve efficiency. Through this collaboration, customers can use features such as “Instant Checkout” to purchase products, replenish household items or plan meals through natural conversation, while Walmart handles backend logistics.
Walmart CEO Doug McMillon said the shift moves beyond the traditional search-and-list model toward a more personalized and contextualized “native AI experience.” By combining OpenAI’s technology with its own AI initiatives, the retailer is transitioning toward an interactive, multimedia future designed to make shopping more intuitive and convenient.
Data Infrastructure as a Barrier to Entry
The success of agentic commerce depends heavily on the quality of an organization’s data infrastructure. Research from MIT Sloan indicates that 82% of executives cite organizational data quality as the primary barrier to achieving generative AI objectives.
To enable accurate AI performance, retailers are implementing data fabrics and ontologies that provide structured representations of data relationships. Case studies show that enriching data with semantic layers can improve AI model accuracy from 16% to 54%. McKinsey research further indicates that retailers that successfully integrate proprietary data into AI systems achieve EBITDA margins up to 25% higher than peers.
To maintain control over how products appear in AI-driven results, retailers are pursuing three main strategies:
-
Web Scraping Optimization: Ensuring websites are structured so AI crawlers can extract accurate data.
-
Model Context Protocols (MCPs): Using API-based frameworks to provide AI platforms with a structured blueprint for data retrieval.
-
Direct Syndication: Delivering highly structured product feeds directly to AI platforms like Perplexity and OpenAI through syndication partners such as Feedonomics.
Conversational Commerce and Business Agents
Google has also introduced “Business Agent,” a tool that allows retailers to deploy virtual sales assistants directly within search results. These agents answer product questions using a brand-specific voice and are currently being piloted by retailers including Lowe’s, Michaels, Poshmark and Reebok. Future updates will allow the agents to be trained on proprietary data and support direct agent-to-agent payments.
In parallel, Google is testing “Direct Offers” within AI Mode. This pilot enables advertisers to present exclusive, real-time discounts to shoppers identified by AI as ready to buy. For instance, a consumer searching for a dining room rug could see relevant product options alongside a limited-time 20% discount from a participating retailer.
Industry experts suggest that by late 2026, commerce will shift from AI-assisted decision-making to a model in which autonomous agents serve as the primary interface for global commerce.