AI agents could drive $6.2 trillion in purchases by 2030, challenging banks to lead security infrastructure

Economic forecast and the new role of payment systems

A Bain & Company analysis predicts that purchases made by artificial intelligence agents could reach up to $6.2 trillion in global sales by 2030. In this model, algorithms not only search for products and compare prices but can also select the most suitable payment method based on costs, speed, and rewards programs.

How payment method selection is changing

While consumers traditionally used the same card out of habit, an AI agent can instantly evaluate all available options and choose the one that best meets the user’s predefined rules. The decision is no longer based solely on personal preference but on technical criteria established at the time of purchase.

The role of banks and the payments industry

According to Luiza Cerri, a partner at Bain & Company, the financial sector must move beyond merely observing technology and take the lead in creating a trust and security infrastructure that validates every algorithm-executed transaction.

The financial sector’s key challenge is not just to keep up with this technology but to lead the creation of a trust and security infrastructure that validates every algorithm-executed transaction.

Banks can monetize identity verification services, security protocols, and proprietary data to enhance protection against fraud in autonomous transactions. Even when interactions occur on third-party platforms, these institutions can generate revenue.

Global collaboration to identify AI agents

In September, Visa, Mastercard, and Ant International announced a partnership to establish common standards for verifying AI agents’ identities and linking them to responsible entities. The initiative aims to answer a fundamental question: who authorized the software to act, and what are the limits of that authorization?

Tests already underway in Brazil

In March 2026, Brazil took a significant step when Banco do Brasil and Visa conducted a transaction initiated by an AI agent in a controlled production environment. The operation used the Visa Intelligent Commerce platform and tokenized credentials, which replace sensitive card data during payment.

This experiment demonstrates how technology can be applied in controlled conditions, but achieving the projected volume still depends on consumer adoption and integration among agents, merchants, and financial institutions.