The AI Shopping Wars Heat Up
How ChatGPT and Amazon Are Redefining Online Commerce
OpenAI launched shopping research in ChatGPT on November 24, 2025, creating a tool that asks users clarifying questions about their needs, searches across retailers, and builds personalized buyer’s guides. The feature uses a specialized version of GPT-5 mini trained with reinforcement learning for shopping tasks. Shortly after this launch, Amazon blocked multiple OpenAI web crawlers from accessing its product pages through updates to its robots.txt file, preventing ChatGPT from reading Amazon’s prices, specifications, and reviews. This move affects how AI shopping tools can access the world’s largest online retailer.
The Clash Between AI Discovery and Retail Control
ChatGPT Shopping
AI Product Research Assistant
VS
Amazon
E-Commerce Platform
The shopping research feature rolled out to all ChatGPT users on free, Go, Plus, and Pro plans, with nearly unlimited usage available through the holiday season. When users ask shopping-related questions, ChatGPT automatically suggests using the research tool, which takes several minutes to generate detailed product comparisons. The system searches for up-to-date pricing, availability, reviews, specifications, and images from publicly available retail sites. Users can guide the research by marking products as “Not interested” or “More like this,” allowing real-time refinement of recommendations.
800M
Weekly Active ChatGPT Users
52%
Product Accuracy Rate
GPT-5
Mini Model Powered
Free
Available to All Plans
Timeline of Events
November 24, 2025
OpenAI introduces shopping research feature powered by a version of GPT-5 mini trained specifically for shopping tasks using reinforcement learning, achieving 52% accuracy on multi-constraint product queries.
Shortly After Launch
Amazon updates its robots.txt file to block ChatGPT-User and OAI-SearchBot crawlers from accessing product pages, joining previously blocked GPTBot used for model training.
November 25, 2025
Perplexity launches its own AI shopping assistant with PayPal integration, allowing users to complete purchases directly within the conversational interface for participating merchants.
Industry Response
Retailers including Walmart, Target, and Etsy see increased referral traffic from ChatGPT and establish partnerships with OpenAI, while Amazon maintains its blocking stance to protect its advertising business.
ChatGPT Shopping Research Performance Analysis
OpenAI trained the shopping research model to read trusted sites, cite reliable sources, and synthesize information across multiple retailers. The company designed the system to be transparent, stating that user chats are never shared with retailers and results are organic based on publicly available retail sites. Merchants who want to appear in shopping research results can follow an allowlisting process through OpenAI. The feature performs particularly well in detail-heavy categories including electronics, beauty, home and garden, kitchen appliances, and sports equipment.
What This Means for the Shopping Ecosystem
For Online Shoppers
The shopping research tool handles complex queries with multiple constraints, such as finding products that meet specific price, feature, and compatibility requirements. However, Amazon’s blocking means the tool cannot access product data from the largest online retailer. OpenAI acknowledges the model may make mistakes about pricing and availability, recommending users verify information on merchant websites before purchasing.
For Retail Competition
Smaller retailers gain potential visibility as AI shopping tools search across multiple sites rather than defaulting to Amazon. Walmart now receives approximately 36% of its referral traffic from ChatGPT, up from 20% in earlier reports. Brands that optimize their sites for AI discoverability through proper structured data may see increased traffic, while those blocking AI agents risk becoming invisible in this emerging channel. Companies like those backed by major tech investors are positioning themselves for this shift.
Technical Implementation
The shopping research model achieved 52% product accuracy compared to 56% for GPT-5-Thinking, 37% for GPT-5-Thinking-mini, and standard ChatGPT Search on difficult product discovery queries with multiple constraints. Product accuracy measures how well recommended products meet user requirements for attributes including price, color, material, and specifications. This specialized training required creating new evaluation benchmarks for complex shopping scenarios. Similar advances in AI model capabilities are reshaping multiple industries.
Business Model Implications
Amazon generates approximately $56 billion annually from advertising, a business dependent on shoppers browsing Amazon.com rather than receiving recommendations through external AI agents. By blocking OpenAI crawlers, Amazon protects this revenue stream while maintaining control over customer relationships and data. However, CEO Andy Jassy stated on earnings calls that Amazon is “having conversations” with AI shopping agents and expects to “find ways to partner” over time. Meanwhile, developments in AI capabilities continue advancing across the industry.
OpenAI is not currently monetizing the shopping research feature through affiliate revenue or advertising. The company stated it sources review sites so users can visit them to read full reviews and potentially use buy buttons. Future updates will include Instant Checkout capabilities, allowing direct purchases within ChatGPT for participating merchants. The feature also integrates with ChatGPT Pulse for Pro users, which can proactively suggest personalized buyer’s guides based on past conversations. Integration efforts across AI platforms, similar to other AI tool developments, are expanding AI’s role in daily tasks.
Perplexity launched its competing shopping assistant on November 25, 2025, one day after OpenAI’s announcement. The Perplexity tool uses the platform’s memory features to tailor recommendations based on what it knows about users’ locations, professions, and preferences. Through integration with PayPal, Perplexity users can complete purchases directly within the conversational interface for merchants accepting PayPal. Google also recently added AI-powered shopping features to AI Mode in Google Search, intensifying competition in this space. The broader advancement of AI models and AI infrastructure investments are enabling these shopping innovations.
The shopping research feature in ChatGPT was discussed, including its technical implementation using a specialized GPT-5 mini model, performance metrics showing 52% accuracy on complex product queries, and availability to all user tiers. Amazon’s blocking of OpenAI web crawlers through robots.txt updates was covered, affecting ChatGPT’s ability to access Amazon product data. The competitive response from Perplexity and partnerships with retailers like Walmart, Target, and Etsy were noted. The business implications related to Amazon’s $56 billion advertising revenue and the shift toward AI-mediated shopping experiences were examined.
