AI Chatbots as Emerging Referral Powerhouses for Retailers

Artificial intelligence chatbots, particularly ChatGPT, have rapidly become significant sources of referral traffic for major retail websites. Retail giants such as Walmart, Target, Etsy, and eBay are witnessing a notable influx of visitors directed from AI-driven chat platforms, while Amazon has taken a contrasting approach by limiting AI access to its product data.

According to recent analytics from Similarweb, ChatGPT accounted for approximately 20% of Walmart’s referral traffic in August 2025, marking a 15% increase from the previous month. Etsy benefits even more, with over one-fifth of its referral visits originating from ChatGPT, while Target and eBay receive nearly 15% and 10%, respectively. Although these referral clicks represent a small fraction-less than 5%-of total site visits, the rapid growth highlights AI’s expanding role in shaping consumer shopping journeys.

Shifting Consumer Behavior: From Search Engines to AI Recommendations

Traditionally, consumers have relied heavily on search engines like Google to discover products. However, AI chatbots are increasingly becoming the go-to resource for product inquiries and recommendations. A recent study by OpenAI’s Economic Research team and Harvard economist David Deming estimates that about 2% of all ChatGPT queries-roughly 50 million daily-are related to shopping. With ChatGPT processing 2.5 billion prompts each day, this represents a substantial volume of purchase intent.

Users frequently request tailored advice, such as “suggest a reliable smartphone under $500” or “compare prices for running shoes,” indicating that AI tools are assuming roles once dominated by traditional search engines. Complementing this, a 2025 survey by Omnisend found that nearly 60% of American consumers have utilized generative AI tools for online shopping assistance, with 25% asserting that ChatGPT outperforms Google in product research.

Amazon’s Strategic Withdrawal and Focus on Proprietary AI

Unlike its competitors, Amazon has actively restricted AI crawlers from accessing its product listings, effectively limiting ChatGPT’s ability to provide real-time Amazon product links. This defensive strategy has led to a nearly 18% decline in ChatGPT-driven referral traffic to Amazon in August 2025, dropping below 3% month-over-month.

Amazon’s marketplace, hosting over 600 million product listings, is the largest e-commerce data repository globally and underpins a $56 billion advertising ecosystem. Allowing third-party AI tools to bypass its storefront could jeopardize both traffic and ad revenue streams. Instead, Amazon is investing heavily in its own AI chatbot, Rufus, which integrates directly into its search interface and features embedded video advertisements. By late 2024, Rufus had fielded over 500 million customer queries, signaling Amazon’s commitment to an in-house AI shopping assistant.

Walmart and Other Retailers Embrace AI-Driven Shopping Agents

In contrast to Amazon’s restrictive stance, Walmart and other retailers like eBay have welcomed AI bots crawling their sites, allowing ChatGPT and similar tools to surface their products. Walmart, with approximately 420 million SKUs, now occupies a significant share of AI chatbot shopping results, effectively becoming a dominant presence on the “agentic shopping shelf.”

Walmart is also developing its own AI shopping agents, designed to assist customers directly on its platform. Hari Vasudev, Walmart U.S. CTO, anticipates the retail industry will eventually adopt standardized protocols enabling third-party AI shopping assistants to interact seamlessly with retailer systems, fostering an open ecosystem for AI-driven commerce.

Industry analysts note that AI systems tend to prioritize well-established, trusted retailers over smaller independent stores, reflecting a bias toward major brands in AI-generated shopping recommendations.

The Future of AI-Driven Commerce: Monetization and Market Impact

Currently, retailers benefit from a surge of complimentary referral traffic from AI chatbots. However, this dynamic is poised to evolve as companies like OpenAI explore integrating payment and checkout functionalities directly within chat interfaces. Such developments could enable users to complete purchases without leaving the AI platform, opening new revenue streams through affiliate commissions, advertising, or transaction fees.

Experts warn that the era of free AI-driven traffic may soon end. Juozas Kaziukėnas, an independent market analyst, emphasizes that as AI shopping tools mature, retailers will likely face new costs associated with AI-generated referrals. This shift has already raised concerns among brand executives about the future economics of AI-powered e-commerce.

Moreover, the immense computational expenses required to sustain AI technologies present a financial challenge. Bain & Company projects that by 2030, global AI infrastructure costs could reach $2 trillion annually, outpacing expected revenues by $800 billion, underscoring the urgency for viable monetization models.

Adapting to the AI Shopping Revolution

Marketing agencies and brands are rapidly adjusting to the AI-driven landscape. For instance, Azoma, a digital marketing firm, reports that clients have experienced a sevenfold increase in website visits originating from ChatGPT since adopting AI-focused strategies. This measurable return on investment highlights the growing importance of optimizing for AI referral traffic.

As AI assistants like OpenAI’s Operator evolve toward fully autonomous shopping capabilities-where bots independently complete purchases on behalf of consumers-the retail sector faces transformative changes. Even if fully autonomous AI shopping remains a future prospect, the current integration of AI-generated product links is already reshaping consumer behavior and retail marketing.

Kaziukėnas summarizes the impact succinctly: “The influence of AI on shopping spans a spectrum-from simple product links in chat responses to potentially fully autonomous purchasing agents. We are just at the beginning, but even this initial phase is profoundly disruptive.”