WEX Uses Agentic AI to Redesign Its Operating Model

AI agents are pushing corporate artificial intelligence beyond copilots that draft or analyze and into systems that execute multi-step work across enterprise software. That shift matters because autonomous action can change payments, compliance and operating decisions, not merely employee productivity. Executives at Visa and FIS say the “agentic enterprise” therefore requires redesigned workflows, unified data foundations, explicit permissions, audit trails and human oversight before companies can translate faster automation into measurable revenue, cost or risk outcomes.

The transition is becoming concrete in financial services. Visa on July 14, 2026, unveiled AI Financial Assistant, with a U.S. financial-institution pilot scheduled for August. FIS on May 4 announced a Financial Crimes AI Agent co-developed with Anthropic, targeting a market where the United Nations estimates $2 trillion in illicit funds move through the global financial system annually and U.S. institutions spend $35 billion to $40 billion a year on anti-money-laundering operations. FIS plans broader availability in the second half of 2026.

Agentic AI is changing how companies use artificial intelligence, moving beyond report generation and passive analysis to systems that proactively identify problems, recommend actions and help carry them out. The technology could shorten management decision cycles and shift organizations from labor-intensive operations toward a model in which humans provide oversight and AI agents work alongside them.

Recent reports suggest AI agents could free up about 40% of employees’ time, allowing them to focus on judgment, creativity and other high-value work. Companies would also no longer need to wait for reports before deciding what to do next. However, the available information does not identify the research organization, publication date, survey sample or investment amount, and more empirical evidence is needed to validate the reported benefits.

For the past two decades, enterprise software has largely relied on SaaS contracts priced by seat or license, with vendors primarily guaranteeing uptime and offering compensation for outages. But Agentic AI can now approve refunds, reconcile accounts and initiate payments, meaning errors can directly cause financial losses. Mayer Brown and Stoel Rives therefore argue that contracts must include human review, audit rights and clearer allocations of liability.

PYMNTS reported on Feb. 24, 2026, that Mayer Brown had said on Feb. 17 that these contracts were shifting toward managed-services or outsourcing structures. On March 19, it also cited Nvidia CEO Jensen Huang as saying at GTC 2026 that SaaS would move toward outcome-based “GaaS” pricing. A survey found that 43% of chief financial officers expected Agentic AI to have a high impact on dynamic budget reallocation, while 47% anticipated a moderate impact. The reports did not disclose actual contract values.