Dal Sahota
Global Director of Trusted Payments, London Stock Exchange Group (LSEG)
Why the emerging world of agentic payments is a matter of trust
Before this year’s World Cup, Dal Sahota, Global Director of Trusted Payments at London Stock Exchange Group (LSEG), saw an impressive agentic payments demonstration at a conference. “The scenario was finding tickets to watch England play, but only in certain cities, on certain days and up to a maximum price,” he remembers.
AI agents benefit consumers and businesses
“The AI agent took that information, searched for available games within that budget and brought back options. The consumer would then select the one they preferred, and within three clicks, the transaction and all the bookings were completed seamlessly,” explains Sahota.
Agentic isn’t just making its presence felt in the consumer space. To improve efficiency and productivity, businesses are starting to use AI agents for regular supplier payments. This type of innovation feels almost inevitable, says Sahota, because AI is evolving so rapidly.
“New technologies are continuously being created to benefit consumers, businesses and economies,” he points out. “It’s only natural that we see greater levels of adoption.”
For agentic payments to go mainstream, users need to have faith in it.
Making innovations safe and more secure
There is, of course, a very large ‘however’ looming over all of this. “When innovation takes the front seat, and consumer protection is not built in from the outset, we can end up in some terrible scenarios,” emphasises Sahota.
“The global financial crisis is a good example. New products created opportunities, but the controls and safeguards around them didn’t always evolve at the same speed. The result was a loss of trust and significant consequences for consumers and businesses alike.”
Questions surrounding trust in AI agents
For agentic payments to go mainstream, users need to have faith in it. That means trusting that AI agents are not susceptible to being hacked by cybercriminals. At present, that confidence is perhaps lagging.
Agents that incorporate robust authentication, verification, guard rails and cyber monitoring controls are likely to drive broader market adoption by increasing trust and reducing risk, says Sahota.
Organisations are also carrying out continuous testing to ensure agents remain secure, operate within defined guardrails and can be traced to a verified source. Key questions include who created the agent and whether that creator has been authenticated as a legitimate person or trusted agent.
All of this means there are ways for agentic payments to achieve consumer and business confidence. “But are safety and security ubiquitous and working hand in hand in the way we all wish to see?” asks Sahota. “I don’t think we’re quite in that space yet.”
Accountability for payments made via AI agents
There are other issues, too. Say you are scammed by a fraudulent AI agent. You might not just lose money on the one transaction you made with it. If it has your financial details, in a worst-case scenario, it could go rogue and begin emptying your bank account. With a non-human in the mix, will current reimbursement schemes cover payments made via agents? There are questions about who’s liable: the bank, service provider or the agent’s author?
Robust controls to prevent consumer and business fraud
Sahota believes regulation is a more effective “stick” in ensuring good governance and protecting consumers than broader guidance is a “carrot.”
“If companies are required by law to have robust controls in place to prevent consumer and business fraud, the outcome would be very different,” he says.
“That’s what we’re seeing in the EU and the UK with security checks such as Confirmation of Payee and Verification of Payee. So, I think we’re going to have more regulatory consumer-driven protection above and beyond reimbursement schemes; and that will work favourably for the market.”
In the meantime, Sahota has this advice for consumers and businesses using agentic payments: “Qualify the risk you’re taking against the payments you’re making — or, rather, the decisions you are asking the agent to make on your behalf. The greater the authority you give an agent, the greater the confidence you need in its decisions.”
