The Battle Over AI Shopping Agents: Data, Deals, and Development Challenges
AI chatbot developers and major retailers are engaged in complex negotiations to establish the foundation for agentic shopping—where AI assistants handle online purchases autonomously. While partnerships between companies like OpenAI, Google, and Walmart signal progress toward this future, significant technical and data-sharing challenges remain unresolved. Current implementations require substantial user oversight and operate slowly, with concerns about accuracy, personalization, and transaction reliability delaying widespread adoption. The industry faces critical questions about data exchange, revenue sharing, and consumer trust as it works toward making AI shopping assistants a practical reality.
The emergence of agentic shopping represents one of the most significant developments in e-commerce, promising a future where AI assistants handle our online purchases autonomously. However, this vision is currently mired in complex negotiations between AI developers and retail giants over data access, revenue sharing, and technical implementation. As companies like OpenAI, Google, and Amazon race to establish partnerships with major retailers, they're confronting fundamental challenges that could determine whether AI shopping agents become a convenience or remain a technological novelty.
The Current State of Agentic Shopping
Despite ambitious announcements and prototype features, today's AI shopping agents remain limited in functionality and reliability. Current implementations require significant user input and operate slowly, often taking considerably longer than manual shopping. In testing, simple tasks like adding items to a cart can take AI agents up to three times longer than human users. These limitations stem from ongoing negotiations about how much product data and chat history need to be exchanged between AI companies and retailers to create effective shopping experiences.
The technical challenges are substantial. AI agents must access real-time pricing, availability, and delivery information while maintaining the personalized, conversational experience that makes chatbots appealing. Retailers, meanwhile, are protective of their customer data and competitive pricing information, creating tension in partnership discussions. As Amazon CEO Andy Jassy noted in recent earnings comments, current agentic shopping experiences lack personalization, shopping history integration, and frequently provide inaccurate delivery estimates and pricing.
Partnership Landscape and Technical Compromises
The partnership ecosystem for agentic shopping is rapidly evolving. OpenAI has partnered with Walmart to enable ChatGPT users to purchase Walmart products directly within chat windows. Both OpenAI and Perplexity have announced deals with payment processors like PayPal and e-commerce platforms like Shopify. Google recently introduced AI agents capable of filling out online checkout forms and calling stores for pricing information.
These partnerships aim to address the technical compromises needed for agentic shopping to succeed. Payment processors like Visa and software startups such as New Generation are attempting to broker solutions that balance retailer needs with AI capabilities. The goal is to create systems where chatbots not only present accurate product information but also consume fewer computing resources when executing orders—potentially boosting profits for both AI companies and retailers if they can establish mutually beneficial terms.
Data and Revenue Sharing Challenges
Money and data represent the core negotiation points in agentic shopping development. AI companies including Opera are seeking revenue shares from sales they facilitate, arguing that they deserve compensation for driving incremental purchases. OpenAI has established a precedent by collecting what it describes as "a small fee" from partners like Etsy for Instant Checkout purchases.
Data sharing presents more complex challenges. Retailers closely guard pricing, availability data, and customer information to maintain competitive advantages. AI companies want to protect conversation histories to preserve the intimate chatbot experience. However, effective shopping agents require real-time information to fulfill user requests, while retailers prefer greater context to develop customer relationships. This tension is evident in current implementations where partners like Expedia receive user IP addresses and relevant chat queries but desire more comprehensive conversation data to better understand customer preferences.
Industry Skepticism and Future Outlook
Despite the hype surrounding agentic shopping, significant industry skepticism remains. Many executives describe current offerings as marketing announcements rather than solid solutions. Some smaller AI companies are deliberately avoiding agentic shopping deals, focusing instead on improving recommendation quality before tackling autonomous purchasing. As one travel industry executive noted, the consequences of AI errors in shopping are substantial—nobody wants to ruin a family vacation because a bot made incorrect booking decisions.
The path forward requires addressing fundamental reliability concerns while establishing fair data and revenue sharing models. Until these challenges are resolved, consumers hoping to fully delegate their shopping to AI this holiday season will likely need to maintain active oversight. The industry's ability to solve these issues will determine whether agentic shopping becomes the next major e-commerce revolution or remains an intriguing but impractical technological concept.
