PGA of America Streamlines Identity Architecture to Manage Production AI Agents

As enterprise artificial intelligence tools expand from basic assistants into autonomous agents that touch core business systems, identity management is becoming a critical security boundary. For the Professional Golfers’ Association of America, managing that shift required dismantling legacy infrastructure rather than adding new administrative layers, according to Chief Technology Officer Kevin Scott.

Speaking on SiliconANGLE Media’s theCUBE broadcast during Okta Inc.’s Oktane conference, Scott detailed how operating with a lean IT team prompted the organization to rely heavily on Okta and Auth0 to secure systems without swelling operational headcount. SiliconANGLE reported on the strategy at the event (https://siliconangle.com/2026/09/24/the-pga-of-america-secures-identity-security-oktane/).

“I think the biggest one is we don’t have much of a team – it’s small, it is lean,” Scott said. “I don’t mean to be reductive, but if we were to spend half of the technology budget on administering systems, none of our customers are ever going to come along and say thank you for that.”

To eliminate fragmentation where identity policies could fail, the PGA of America phased out physical data centers, virtual private networks, Active Directory, and traditional corporate network infrastructure. Scott explained that managing user personas across multiple disparate systems introduces risk, making structural reduction the most practical defense.

The introduction of AI agents—which can navigate enterprise tools on behalf of staff—introduced potential sprawl. Rather than permitting independent agents to multiply, the PGA of America directs employees toward a unified AI platform, wrapping identity controls around technologies such as the Model Context Protocol (MCP) and agent gateways to authenticate individual users.

“We try to make ours the best option,” Scott said. “Use gravity to create the right defaults for people. That has allowed us to … centralize things. We don’t have a proliferation of agents around. We kind of know where they’re going to be running it from and we can authenticate … each independent user in all the ways that Okta’s been talking about this week.”

The deployment of AI has also shifted internal decision-making. Software procurement and architecture choices are no longer isolated within IT, requiring joint review across legal, human resources, finance, and engineering stakeholders. The organization routes new use cases through an internal council with human review, while maintaining enough hands-on development experience to evaluate vendor tools and make informed build-versus-buy decisions.