The new operating system, agentOS, is designed to help financial institutions deploy, manage and scale AI agents across their workflows. agentOS is running in beta with two financial institutions and is expected to be widely available by August, Fiserv said in a Thursday (May 14) press release.
agentOS is built to operate across Fiserv’s FinTech platforms and is supported by an agent marketplace that allows financial institutions to access Fiserv-built agents, build their own or deploy third-party agents within the system’s controlled architecture.
“The pilots are already proving it works — delivering measurable, gains today, with six financial institutions co-developing the next wave with us,” Fiserv Co-president Dhivya Suryadevara said in the release.
Fiserv’s collaboration with OpenAI, which was announced in another Thursday press release, will see the two companies work together to build agents on agentOS.
In addition, Fiserv and OpenAI will explore how financial institutions can modernize faster while reducing risk and without disrupting their day-to-day operations; how to bring banking-specific AI capabilities to Fiserv platforms; and how to use AI to advance the cybersecurity capabilities Fiserv already delivers to financial institutions.
Advertisement: Scroll to Continue
Suryadevara said in the release that this collaboration “lets us bring frontier AI directly into our platforms and products and how we serve our clients — so our clients see real, tangible differences in their day-to-day operations.”
Ashley Kramer, vice president, enterprise at OpenAI, said the collaboration will help financial institutions “modernize faster, operate more intelligently and deliver better experiences for their customers.”
PYMNTS reported May 6 that OpenAI, Anthropic, Google and Microsoft are among the AI companies that are competing for enterprise financial services business. AI firms are attempting to become integrated layers inside banking infrastructure, compliance operations, fraud systems and treasury management.
Fiserv executives said during a May 5 earnings call that in its own operations, the company is leaning heavily into AI to improve client support. They reported that the time required to resolve client inquiries has dropped 27% year over year, with “high impact client incidents” down 60%.
