TL;DR: Despite partnerships between AI companies (OpenAI, Google, Perplexity) and retailers (Walmart, Shopify, Etsy), agentic shopping agents face significant implementation barriers. Ongoing negotiations centre on data sharing, liability for mistakes, and performance issues—Opera’s AI agent took 45 seconds to add eggs to an Amazon cart versus 15 seconds manually. McKinsey projects $1 trillion in US agentic shopping sales by 2030, but current features require substantial user input and operate with limited functionality.
The Promise Versus the Reality
OpenAI’s ChatGPT now enables instant checkout for Etsy products directly within the app, representing early steps toward agentic shopping. Yet according to executives at seven tech and e-commerce companies, AI chatbot developers and major retail partners remain locked in negotiations over fundamental issues: limiting costly agent errors, determining data exchange requirements, and establishing revenue-sharing models.
Current market features reflect these constraints. Users must provide significant input, systems operate slowly, and functionality covers only limited product catalogues. Consumers hoping to delegate shopping chores to automation this holiday season face a gap between marketing announcements and functional reality.
Talia Goldberg, partner at Bessemer Venture Partners (investor in Perplexity and Fal), notes she hasn’t yet experienced a “super magical agentic experience in commerce,” highlighting big questions around creating truly functional experiences.
Market Potential and Partnership Activity
Consumer surveys indicate 60% of US consumers plan to use AI for shopping assistance, with 20% willing to let AI agents fully handle everyday purchases. McKinsey estimates agentic shopping could generate up to $1 trillion in US sales by 2030.
Recent partnerships aim to realise this potential. OpenAI partnered with Walmart for in-chat purchasing, whilst both OpenAI and Perplexity announced deals with PayPal and Shopify. Google introduced AI agents capable of filling online checkout forms and calling stores for pricing information.
Expedia’s ChatGPT app demonstrates promise, providing real-time flight and hotel pricing in response to queries. Whilst users must manually complete bookings, the feature generates higher sales than anticipated. Clayton Nelson, Expedia’s vice president for strategic alliances, states: “That means there’s something in these tools that works.”
The Data and Money Negotiations
Financial arrangements appear straightforward—AI companies seek sales commissions for facilitating purchases. OpenAI collects “a small fee” from partners like Etsy for Instant Checkout transactions. Opera executives similarly expect compensation for adding “incrementality” to sales.
Data sharing proves more complex. Retailers guard pricing, availability, and customer data to maintain competitive advantages. AI companies protect conversation histories to preserve chatbot intimacy. Yet chatbots require real-time information whilst retail brands want greater shopper context.
OpenAI’s apps share users’ IP addresses and relevant chat queries with partners like Expedia. Nelson wants expanded access—with user consent—to full conversation context: “I want to know, are the guests friends? Have they traveled before? Do they have other things that they like?”
Some discussions have deteriorated. Amazon sued Perplexity this month for allegedly using AI agents to interfere with Amazon’s advertising and Prime subscription businesses. Perplexity stated it will contest the action.
Performance and Trust Barriers
Amazon CEO Andy Jassy delivered frank criticism of current agentic shopping implementations during an earnings call: “The customer experience is not good. There’s no personalisation. There’s no shopping history. The delivery estimates are frequently wrong. The prices are often wrong.”
Performance testing supports these concerns. Opera’s AI agent required 45 seconds to add eggs to an Amazon cart—triple the manual time on Amazon’s shopping app.
Payment processors like Visa and software startups such as New Generation (which helps stores develop chatbot partnerships) are attempting to broker technical compromises. New Generation CEO Adam Behrens suggests service providers may earn retailer trust faster than tech giants, referencing resistance to social commerce platforms like TikTok and Instagram.
The accuracy stakes are particularly high for complex purchases. Nelson emphasises: “No one wants to mess up their vacation for their entire family because a bot went left instead of right, or didn’t follow the specific prompt that was given.”
Amazon’s own Buy for Me feature—using agentic AI to complete purchases on third-party sites when Amazon lacks stock—illustrates both opportunity and caution. The tool adds items to carts whilst users check out with Amazon payment details, but third-party stores can request exclusion.
The Path Forward
A California clothing retailer executive notes that whilst chatbots drive significant traffic, current opportunities feel undeveloped: “Up to today, no one has a solid solution. Everyone is just making marketing announcements.”
Some AI companies are deferring agent shopping entirely. Archit Karandikar, CEO of travel planning chatbot Airial, states that generating useful recommendations represents sufficient challenge without adding purchase agency: “You can’t be spending someone’s money without being sure you’re making the right transactions.”
This holiday season, chatbots may assist with gift selection and cart management—but humans will retain control. Perhaps next year, users can blame bots for unpopular gifts.
Source: WIRED