Mastercard’s report suggests agentic AI could turn the logic and rules of payments on its head.
Mastercard has predicted that one in 10 consumers will use AI assistants to shop and pay by 2030.
But perhaps more consequential for the payments industry is the death of the traditional “top of wallet” payment card.
Issuers have competed for years to become the card a consumer takes from their wallet at checkout, or more recently saves as their default in their mobile devices, persuading users with rewards, sign up bonuses and brand loyalty.
According to Mastercard’s new report, ‘A Short History of the Future of Shopping and Payments‘, AI agents could strip that behaviour by evaluating payment options in real-time, selecting whichever best matches a consumer’s instructions.
The report delves into how agentic AI could impact shopping, retail and payments as software stops answering questions and starts acting on behalf of consumers.
Theodora Lau, author and commentator on the future of banking and finance and contributor to the Mastercard report, brought up how agentic shopping and payments could end the fight for top of the wallet as we know it.
She explained that the top of wallet status is partly achieved through habit, with a consumer using the same card because of cashback, rewards, familiarity or because it is already saved in an app. However, an AI agent doesn’t possesses the same attachment.
Lau notes that it would compare factors including interest rates, cashback, foreign transaction fees, fraud liability and compliance status before choosing the payment method that provides the strongest outcome.
“Top of wallet becomes more like a scorecard than a habit formed over years,” she wrote.
By 2030, Mastercard predicts AI agents will reoptimise consumers’ wallets according to objectives and constraints set by the user, edging it closer to the role of a personal “financial chief of staff” than a checkout tool.
Banks and payments providers will therefore be faced with different competitive challenges as attractive branding or app engagement will likely matter less if the purchasing decision is made by software.
Payments infrastructure made for machines
Card selection won’t be the only area that institutions will have to rethink, with Lau stating that the majority of current banking and payments infrastructure assumes there is a human at the other end of a transaction.
AI agents don’t respond to push notifications in the same way as consumers or telephone customer service when something goes wrong, meaning that agentic commerce needs infrastructure that can authenticate non-human actors, establish what they have permission to do and know who’s responsible when a transaction fails.
The Mastercard report predicts that Know Your Agent (KYA) could become an important compliance concept, forecasting that at least three G20 financial regulators will issue formal guidance by 2030 covering how AI agents should be registered, authorised, monitored and held accountable.
Payment Expert heard similar predictions from Garima Chaudhary, VP of Financial Crime and Compliance AI at Thetaray during Money20/20 Europe in June.
“Agentic payments completely shift the assumptions anti-money laundering systems are designed on,” she said. “Instead of thinking what a normal human behaviour is, we need to think about what a dedicated normal behaviour is.”
Mastercard on how trust will change
Research referenced in the report found roughly one in five UK adults would be likely to use AI acting autonomously within goals they set, with trust and control being their main concerns.
Trust has played an important role in payments from the beginning of time. However, reputation or familiarity with a brand isn’t likely to play as much of a role because, unlike humans, AI agents won’t be influenced by emotional triggers.
The report predicts that by 2030, agentic transactions will carry machine readable liability warranties setting out who is accountable if a payment goes wrong.
“The institutions that will win the next decade will be those that can foster trust in AI and close the trust gap, so that consumers will be more willing to let agents act on their behalf – delegating tasks without losing the wheel,” Lau wrote.
“In the agentic world, trust is not a sentiment, but a property – with machine-readable terms.”
The start of a different logic
In the report’s conclusion, Magnus Lindkvist, a futurologist and trendspotter, describes agentic commerce as a potential “Kansas Moment”, noting it as the point when familiar behaviour starts operating according to unfamiliar rules.
“The technology matters, but what matters more is the subtle psychological shift it triggers: the realisation that buying, choosing, and paying are about to work according to a different logic than the one we have always known.”
He explained that there will be naysayers, pointing to people believing that people would never “fly like birds” before the invention of aeroplanes, and once they see it is possible, they will question if it’s useful until they finally accept it as a normality.
“Agentic AI will follow the same path. One day, hearing an AI say, ‘You’ve told me impulse purchases are off-limits. If you still want it tomorrow, I’ll order it,’ will simply feel like good service,” he concluded
“But the future of commerce will not be defined by which agents are fastest, smartest or cheapest. It will be defined by which systems people trust enough to delegate to. The technologies that succeed will be those that combine convenience with confidence, automation with accountability, and intelligence with human agency.”
Want to find out how AI can help your business in practical terms? Get yourself to the AI Academy at SBC Summit Lisbon on Tuesday 29 September. Get yourself an Education+ pass by visiting https://sbcevents.com/sbc-summit/ai-academy/
