Mastercard: AI agents enter the mainstream of shopping by 2030

Mastercard reckons AI agents enter the mainstream by 2030, with a new report predicting that more than a tenth of online purchases will be completed without a human clicking anything. Futurist Patrick Dixon, one of the contributors, expects 300 million shoppers to lean on such tools routinely by then.

The projections blend Mastercard's own research with input from futurists across the US, Europe and Asia. Dixon, who has spent three decades advising corporations on strategy and risk, cautions that uptake will be uneven: markets with deep e-commerce habits will move faster than those where online retail is still thin. Britain books 28% of its retail sales over the internet, against 14% in France and 16% in Germany, while Colombia sits at 7% and South Africa at 10%.

Teenagers are ahead of their parents. A survey commissioned by Mastercard of 26,000 parents and 13-to-18-year-olds across 13 countries, fielded in mid-2026, found adolescents using AI for buying decisions at about twice the adult rate. Some 27% would hand a fully automated assistant the job of recommending, shortlisting and paying, versus 16% of parents; 18% of teens already lean on such tools weekly for discounts, against 10% of parents. Trust is shifting too, with 31% of teens preferring a machine's product pick to a friend's and 23% rating it above their own parents' advice.

Where the stakes rise is money movement. Theodora Lau of Unconventional Ventures expects verifiable consumer intent — a framework Mastercard developed with Google and unveiled in March 2026 that ties identity, intent and action into one privacy-preserving record — to become a cornerstone of financial services, alongside agents that manage wallets against each customer's goals, preferences and limits. Crypto wallets got there first: Coinbase launched Agentic Wallets in February, and Consensys added MetaMask Agent Wallet in June, defaulting to a Guard Mode with daily caps, allowlisted protocols and two-factor checks beyond policy limits, plus an opt-in Beast Mode with fewer interruptions. Governance is the unfinished part — verification of identity, authority and accountability, with fraud, privacy and regulatory safeguards, has to mature at the same pace.

Why it matters for Singapore: Singapore's banks and fintechs are already deep into AI for fraud screening and advice, and MAS has set clear expectations on AI governance and risk management. Agentic commerce raises the bar: with heavy card and wallet usage plus PayNow rails, agents here could transact at scale fast, so verifiable authorisation and audit trails matter early. Whether the city-state keeps its reputation as a trusted payments hub will hinge on keeping those agent permissions auditable and consumers protected.