As agent purchases become more common, how do you counter purchases made by your AI?
Companies are preparing for a near future where consumers will allow AI agents to make purchases on their behalf.
Research has found that many European consumers are already using AI to help shape their purchasing decisions, but not at the checkout, where money passes hands – although that could change, and soon.
‘Agenttic commerce’ is seen as a natural consequence of AI-powered search, which already makes up more than half of the world’s search engine volume. McKinsey’s trend analysis finds that this number could rise significantly in the coming years.
McKinsey found that by 2030, agent trade could reach $5trn globally. But while Morgan Stanley earlier this year noted that only 1pc of consumers choose the agent route, new research elsewhere finds that AI agents could make up a large proportion of the customers a business receives in the coming years.
Behind the scenes, the infrastructure works to make agent trading possible at fintechs like Revolut, Stripe, Visa, Mastercard and PayPal. More are expected to follow.
Did you mean to buy that?
A growing number of users say they can trust AI systems to place orders and make payments on their behalf. But such a combination of trust and automation will eventually create a whole new class of consumer disputes that companies have yet to come before, says Monica Eaton, founder and CEO of Chargebacks 911.
“The commercial infrastructure is being built quickly, but defenses need to evolve at the same pace,” he said.
In the era of agent commerce, both customers and businesses will find it difficult to explain the intent – or lack thereof – when purchases are made by AI agents. It’s easy to determine intent when people make a deliberate choice to push ‘buy’, but agency trading takes that time away from the activity. And at the moment, there aren’t many ways to counter AI-driven shopping, Eaton notes.
“Many customers do not have access to detailed records of the instructions they gave, the permissions that are in place, or how the agent reached their decision. In most cases, transactions are technically authorized, making it difficult to challenge,” he adds.
To solve this, platforms need to prioritize transparency before the transaction takes place. The AI agent in question must be able to demonstrate what it is going to do and why, and confirm that it has the customer’s authority before moving forward with the task. An audit trail of agent purchases will provide an additional layer of protection, Eaton said.
Meanwhile, clear consent structures that define where and what agents can buy, and how much they can spend, will also protect customers.
This may only work in the short term, says Eaton. Long-term protection will include platforms that provide transparency and access to transaction logs, while dispute processes will need to evolve to recognize when an agent’s decision is inconsistent with the customer’s intent.
Shift to commitment
This new category of shopping dispute lies somewhere between fraud and ‘buyer’s remorse’, and current systems are not equipped to deal with this confusion, says Eaton.
“In agent environments, platforms need to take greater responsibility for how instructions are captured, interpreted and used”, and sellers should not be expected to automatically take on this responsibility, he explains.
In addition, if effective frameworks are not built in advance, customers may end up in a situation where they argue against the AI customer service bottleneck regarding unauthorized purchases made by the AI personal agent.
There is still time to get ahead of this, but the window is narrowing, Eaton said. “Businesses need to treat e-commerce as a completely different environment, not just a flash version of existing e-commerce.”
It’s important not to wait for the law to reach you, warns Eaton. “Businesses that build commercial trust early will be in a much stronger position than those that react later.
“As for the future of customer service, it doesn’t have to be AI versus AI. The key is to keep the human at the center of the process. Agency trading must reflect and support human intent. If that principle is lost, trust will follow.”
Don’t miss out on the information you need to succeed. Sign up for Daily BriefSilicon Republic’s digest of must-know sci-tech news.
