AI Shopping Agents Face Major Hurdles Despite Growing Demand

AI shopping agents are advancing rapidly, but significant barriers—including security protocols, liability concerns, and a lack of industry standards—remain unresolved, according to industry leaders. Matt Maher, founder of M7 Innovations, a tech research firm, highlighted that retailers and platforms have yet to establish frameworks allowing agents to complete purchases, citing “security protocols, a lack of agentic commerce standards, and policies blocking third-party agents” as key obstacles.

Why Are AI Shopping Agents Struggling to Complete Purchases?

Despite widespread use of AI for product discovery, agents cannot reliably execute transactions due to technical and regulatory challenges. Melissa Bridgeford, CEO of Wizard Commerce, noted that even for basic tasks like recommending ski gloves, models like OpenAI’s ChatGPT provide specific product suggestions only 9% of the time. “OpenAI fumbled its initial efforts to build a commerce platform, pivoting away from features like Instant Checkout,” she said, adding that retailers such as Walmart abandoned partnerships after the shift.

Courtney Robinson, head of policy at Akoya, emphasized that liability remains a critical roadblock. “If an AI agent makes an unintended purchase, who is responsible? There are no standards for where liability sits,” she explained. This ambiguity has led companies to rely on terms and conditions to limit their exposure, but experts like Maher argue such measures may not shield them from “perceptual liability” if customers demand refunds.

How Do Security Risks Threaten Agentic Commerce?

Cybersecurity concerns are escalating as AI agents expand the attack surface for fraud. Norman Menz, CEO of Flare, a cybersecurity firm, warned that agents could be hijacked by bad actors to make fraudulent purchases or used to create fake identities. “The online fraud problem will only grow exponentially with agents,” he said, citing the risk of stolen credit card data and identity theft.

Adam Winnick, CEO of Finality, a blockchain company, proposed open-source standards for agent verification and transaction monitoring. “Blockchain could help track agent permissions and ensure they act on behalf of legitimate users,” he said. Ben Leventhal, founder of Blackbird Labs, echoed this, noting that identity verification remains a hurdle, even in sectors like dining where payments are handled in person.

What Role Will Blockchain Play in Solving These Challenges?

Blockchain technology is emerging as a potential solution for securing AI-driven commerce. Winnick argued that decentralized systems could create “an identity payload” to verify agent authority, while Leventhal suggested existing identity management firms might develop alternatives. However, experts agree that standardization will take years, leaving merchants to bear the risk of fraud in the interim.

“Merchants already handle ‘card not present’ fraud in e-commerce,” Leventhal said. “They’ll likely absorb the same risks with AI agents until regulations catch up.” Despite these challenges, he remains optimistic about the future. “Innovators will find killer use cases, and once software becomes ‘magical,’ adoption is inevitable,” he added.

Key Takeaways

• AI shopping agents face hurdles in executing purchases due to security, liability, and standardization issues.
• 9% of AI product recommendations fail to provide specific options, according to Wizard Commerce.
• Liability for unintended purchases remains undefined, with no clear regulatory framework.
• Blockchain and open-source standards are seen as potential solutions for fraud and identity verification.
• Merchants may shoulder fraud risks similar to current e-commerce practices until regulations evolve.

Frequently Asked Questions

What is agentic commerce? Agentic commerce refers to AI systems that autonomously perform tasks like product discovery, price comparisons, and purchases on behalf of users.

Why can’t AI agents complete purchases now? Security protocols, lack of industry standards, and unclear liability rules prevent agents from executing transactions. Retailers also fear fraud and legal exposure.

How might blockchain help? Blockchain could enable secure identity verification and transaction tracking, ensuring agents act within authorized parameters.