Why AI Shopping Agents Are Struggling to Take Off
AI shopping agents face significant hurdles despite growing consumer interest, according to industry leaders. Matt Maher, founder of M7 Innovations, highlighted that security protocols, lack of agentic commerce standards, and retailer policies blocking third-party agents are major barriers. Melissa Bridgeford, CEO of Wizard Commerce, noted that even basic product discovery via AI models like ChatGPT fails 91% of the time, with only 9% of responses offering specific recommendations.
OpenAI’s decision to abandon its Instant Checkout feature further complicates adoption. Retailers like Walmart exited partnerships after the shift, according to Bridgeford. “The industry is moving toward agentic commerce, but unresolved issues like fraud, refunds, and returns remain critical roadblocks,” she said.
What Security Challenges Are Exacerbating the Problem?
Norman Menz, CEO of cybersecurity firm Flare, warned that AI agents could “magnify the online fraud problem exponentially.” He cited risks such as bad actors hijacking legitimate agents or creating fraudulent ones using stolen identities. “The attack surface keeps expanding,” Menz said, emphasizing the need for robust identity verification systems.
Adam Winnick of Finality, a blockchain-based platform, argued that open-source standards for monitoring AI agents are essential. “Blockchain could help verify that agents act on behalf of their owners, but no single solution exists yet,” he said.
How Fraud and Liability Risks Are Holding Back Agentic Commerce
Courtney Robinson of Akoya, an open finance platform, stressed that liability in agentic commerce remains “wide open.” If an AI agent makes an unintended purchase, who is responsible? “There are no standards,” Robinson said. Matt Maher added that merchants face “perceptual liability,” where customers demand refunds even if terms and conditions attempt to shift responsibility.
For example, Maher explained, a customer might complain to Gap after an AI agent mistakenly buys a blazer, regardless of legal fine print. “Loyalty matters,” he said. “Customers expect refunds when something goes wrong.”
Can Blockchain Solve Identity and Fraud Issues?
Ben Leventhal of Blackbird Labs, a blockchain-based dining rewards program, believes blockchain could address identity verification. “An identity payload would allow agents to prove they’re acting on behalf of a user,” he said. Leventhal’s company is close to enabling AI agents to book restaurant reservations, though payment fraud is less of a concern in the dining sector.
Winnick acknowledged that existing identity management firms might develop solutions, but “new systems will be needed for AI agents to operate securely,” he said. “Blockchain is one possibility, but not the only one.”
What’s Next for Agentic Commerce? Experts Weigh In
Despite challenges, some see potential. Leventhal predicted that “magical” AI shopping experiences will emerge as developers refine the technology. “The clunkiness will fade, and adoption will follow,” he said. However, he warned that merchants will likely bear the risk of fraud, similar to current “card not present” transactions.
Norman Menz echoed the urgency: “Consumers want AI agents now, but standards take years to form. There will be a demand before solutions exist.”
Did you know?
Only 9% of AI models like ChatGPT provide specific product recommendations when asked about items like ski gloves, according to Melissa Bridgeford of Wizard Commerce.
FAQ: Common Questions About AI Shopping Agents
What is agentic commerce?
Agentic commerce refers to AI agents performing tasks like product discovery, purchasing, or reservations on behalf of users. It relies on standards for security, liability, and identity verification.
Why aren’t AI shopping agents widely used yet?
Barriers include security risks, lack of industry standards, retailer resistance, and unresolved liability issues. Fraud and refund processes also remain untested at scale.
How could blockchain help?
Blockchain could enable identity verification and transaction monitoring for AI agents, ensuring they act within user permissions. However, no universal system exists yet.
Pro tip
Stay informed about AI shopping agent developments by following updates from companies like M7 Innovations, Wizard Commerce, and blockchain startups like Finality.
Call to Action
What do you think the future of AI shopping agents looks like? Share your thoughts in the comments or explore our related articles on AI innovation and e-commerce trends. Stay ahead of the curve by subscribing to our newsletter for more insights.
