Visa and Akamai move to tackle fraud risks in AI-driven payments 

Securing Payments in an Age of AI

Visa and Akamai have announced a new partnership aimed at strengthening security around agentic commerce — transactions carried out by AI shopping agents on behalf of consumers. The focus is on reducing fraud risks as automated payments become more common in the UK and beyond.

By combining Visa’s Trusted Agent Protocol with Akamai’s edge security and behavioural intelligence, the companies say they can better verify who - or what - is making a payment, and ensure it reflects a genuine consumer’s intent.

Addressing the Identity Problem Behind AI Payments

One of the biggest fraud challenges in agentic commerce is identity. Merchants need confidence that an AI agent is authorised, and that it is acting within the limits set by a real person.

The joint approach is designed to authenticate AI agents, link them securely to the individuals who approved them, and preserve payment intent as transactions move from consumer to merchant. This helps reduce the risk of unauthorised spending, spoofed agents or manipulated transactions.

A Growing Concern for UK Merchants

The collaboration comes amid rapid growth in AI-driven shopping and a parallel surge in automated fraud. Research from Akamai shows that AI-powered bot activity has increased dramatically, with billions of automated requests now hitting online platforms.

For UK retailers, this kind of activity is more than a technical nuisance. Malicious bots can distort data, scrape content, enable fraud and undermine trust in digital commerce. Payments are a particularly attractive target, making stronger controls essential as AI agents begin to transact autonomously.

A Layered Defence Against Fraud

Visa and Akamai describe their solution as a layered defence. Payment-network rules and authentication protocols are combined with real-time detection at the network edge, behavioural analysis and threat intelligence.

Together, these measures aim to solve two core fraud problems: telling the difference between legitimate AI agents and malicious bots, and ensuring that payments follow the consumer’s original instructions. However, both companies acknowledge that technology alone is not enough.

What Still Needs to Happen

While the partnership is intended to accelerate trust in agent-led payments and reduce friction for consumers, significant work remains for merchants and payment providers. Fraud controls will need to be supported by clear policies, continuous monitoring and updated risk-management practices.

As AI agents become more capable and more widespread, the industry is likely to need shared standards, wider use of tokenisation and ongoing intelligence sharing to stay ahead of evolving fraud tactics.

Sourced by Noah Wire.

Read more in our Knowledge Hub.