Mastercard Adapts Fraud Systems for Bot-Driven Commerce

Mastercard is recalibrating its fraud detection systems, built over decades to stop bots, as automated agents are increasingly becoming the buyers. Greg Ulrich, the company's chief AI and data officer, stated that the system must now shift from merely blocking bot activity to enabling it.

"We've built a bunch of risk rules over time that were intended to stop a bot from transacting," Ulrich explained at the VB Transform 2026 event. "Now we need to enable the bot to transact, so that requires a change to our risk framework and our risk rules."

The company processes 175 billion transactions annually, assessing fraud risk in less than a tenth of a second for each. New technologies, including generative AI, allow for the incorporation of more data and context, leading to a 300–400% increase in identifying high-risk fraudulent transactions without adding friction or false positives for consumers. Mastercard's Safety Net system has already blocked over 70 billion fraudulent transactions.

Ulrich emphasized that trust is the key to scaling AI. Mastercard has developed a five-layer system to ensure trust in agent-initiated transactions. These layers cover identity verification (including a "Know Your Agent" principle), verifiable intent (a tamper-proof record of instructions), controls (defining purchase limits and constraints), execution (via Mastercard Agent Pay), and intelligence (incorporating risk rules and permissioned insights).

Mastercard sees the largest growth opportunity for AI-driven commerce in business-to-business procurement. For instance, a manufacturer could use an agent to autonomously manage inventory levels and replenish stock within budget and approved supplier lists. This requires a trusted infrastructure that facilitates communication between various agents—procurement, supplier, and banking agents—to enable autonomous operations.