Behind the Scenes: The High-Stakes Negotiations Shaping Agentic Shopping's Future

Imagine a world where your AI chatbot not only recommends the perfect gift but also adds it to your cart, processes payment, and arranges delivery—all without you lifting a finger. This vision of 'agentic shopping' is no longer science fiction; it's the focus of intense dealmaking between AI giants and ecommerce behemoths. Yet, as WIRED reports, the path to this automated future is fraught with challenges, from preventing costly mistakes to negotiating data access and revenue shares.

The allure of agentic shopping lies in its potential to transform consumer behavior. Surveys indicate that 60% of US consumers plan to leverage AI for holiday shopping this year, with 20% willing to let agents handle purchases entirely. McKinsey projects a staggering $1 trillion in US sales driven by these tools by 2030. But current implementations, like OpenAI's Instant Checkout with Etsy, still demand significant user input and operate at a snail's pace for limited items.

The Technical and Trust Barriers

At the heart of these delays are ongoing negotiations over how AI agents interact with retail systems. Executives from seven tech and ecommerce firms told WIRED that key issues include limiting agent errors, minimizing data exchanges, and ensuring real-time accuracy. For instance, OpenAI's partnership with Walmart aims to enable in-chat purchases, while deals with PayPal and Shopify promise broader integration. Google's recent AI agents can even fill out checkout forms or call stores for pricing info.

Yet, prototypes reveal the gaps. Expedia's ChatGPT app delivers real-time flight and hotel data, boosting sales beyond expectations, but users must still manually book. As Expedia VP Clayton Nelson notes, "That means there's something in these tools that works," but full agentic booking is held back by the risk of errors. "No one wants to mess up their vacation for their entire family because a bot went left instead of right."

Retailers are wary, too. They guard pricing, availability, and customer data fiercely to maintain competitive edges. AI firms, meanwhile, seek to protect chat histories for personalized experiences. OpenAI shares users' IP addresses and queries with partners like Expedia, but retailers crave deeper context—like past travel preferences—to build lasting relationships.

Revenue Models and Legal Friction

Money is another sticking point. AI companies like Opera seek commissions on facilitated sales, arguing it's fair compensation for added value. OpenAI charges a 'small fee' to partners like Etsy for Instant Checkout. But tensions flare: Amazon recently sued Perplexity, alleging its agents interfere with ad sales and Prime subscriptions. Perplexity vows to fight back, highlighting the cutthroat nature of these integrations.

Amazon, ever the strategist, is testing its own 'Buy for Me' feature, which uses AI to shop third-party sites when items are out of stock. Users check out via Amazon, shielding their data from other retailers. CEO Andy Jassy has publicly critiqued competitors' agents for lacking personalization and accuracy, calling the current experience subpar.

A Cautious Path Forward

Smaller players are proceeding with caution. CoInvent AI's Airial chatbot focuses on recommendations and affiliate commissions rather than direct purchases, citing the technology's immaturity. "You can’t be spending someone’s money without being sure you’re making the right transactions," says CEO Archit Karandikar.

Payment processors like Visa and startups like New Generation are brokering compromises to build trust. As New Generation's Adam Behrens puts it, neutral service providers can earn retailers' confidence faster than AI giants. This holiday season, expect AI to assist with gift selection and cart additions, but humans will remain firmly in control.

The promise of agentic shopping could redefine ecommerce, blending conversation with commerce in ways that delight users and drive profits. But as negotiations continue, it's clear that getting there requires more than code—it's about forging alliances in an industry where trust, data, and dollars are the real currency. Next year, we might finally see bots taking the blame for bad gifts, but for now, the human touch endures.