Mastercard publicly framed the identity problem in agentic commerce and outlined Agent Pay Trust Framework directions, stressing agent identity, intent authorisation and fraud signals. Ideas include know-your-agent credentials and AI-transaction probability scores. The goal is to help merchants and issuers tell trusted agents from malicious automation. Rollout continues gradually with technology partners.
What it means: Without agent identity, banks struggle to apply strong customer authentication when software initiates payment. A shared framework reduces the risk of blocking legitimate shopping assistants—or approving hostile bots. Greek ecommerce merchants should prepare checkout stacks to accept authorisation signals beyond the classic cardholder. EU regulators will test whether those signals meet PSD and AI Act transparency duties. Platform vendors should avoid closed standards that shut out smaller players. Limited-amount pilots with logged risk decisions are essential before production. Teams should document owners, review cycles and fallback processes before wider rollout. Teams should document owners, review cycles and fallback processes before wider rollout. Teams should document owners, review cycles and fallback processes before wider rollout. Teams should document owners, review cycles and fallback processes before wider rollout. Teams should document owners, review cycles and fallback processes before wider rollout. Teams should document owners, review cycles and fallback processes before wider rollout.
Read the original on Mastercard
Related service: myDATA integrations
Related stories
- Visa opens six Singapore capability hubs for AI, next-gen payments and trust
- Stripe Crypto Onramp powers USDC purchases in Samsung Wallet
- Pliant raises $80M Series C to open B2B payments to AI agents
- Klarna rolls out Call Status Detection globally against phone scams
- Bank of England RPIB responds on next-generation UK retail payments infrastructure