🔥 Top Story
“Bring it to our shop”: Workday’s pitch for keeping AI agents close to your most valuable data
Workday, the HR and payroll SaaS platform, is making the case that enterprises should keep AI agents close to their most sensitive workforce data rather than exporting that data to external AI systems — a pitch the headline frames as "bring it to our shop." In a piece by The New Stack's Frederic Lardinois, Workday CTO Gabe Monroy argues that AI guardrails belong inside the inference engine itself, not bolted on after the fact, because HR and payroll data can't tolerate "good enough" results. The core idea is data gravity: by running agentic workloads where the authoritative employee and pay records already live, Workday positions itself as the governed execution environment for those agents. The discussion touches on AI agents, AI infrastructure, and the Model Context Protocol (MCP) as the connective tissue between agents and enterprise data. Workday has been investing in AI and agents for some time, and this framing emphasizes governance, accuracy, and trust over raw model capability. The takeaway is a strategy that minimizes movement of regulated, high-stakes data while still enabling agent automation.
💡 Why it matters: For platform and security teams, the pitch reframes agent architecture around data gravity — keep regulated HR/payroll data in place and enforce guardrails at the inference layer rather than shipping sensitive records out to external models.
🔗 Read more · The New Stack
DevOps & Infrastructure
Okta is the first to bring AI agent governance inside FedRAMP boundaries
The New Stack
Okta has made "Okta for AI Agents — Core" generally available for FedRAMP- and HIPAA-regulated environments, which it describes as an industry first for bringing AI agent governance inside those compliance boundaries. The product extends AI agent lifecycle management — the provisioning, authentication, authorization, and deprovisioning of agent (non-human) identities — into the same controlled environments that government agencies and healthcare organizations already rely on to manage human identities. In practice, regulated buyers can govern autonomous agents under the FedRAMP and HIPAA controls they are already audited against, instead of standing up a separate, unvetted system. The New Stack's Darryl K. Taft reports that Okta positions this as a milestone for treating agents as first-class identities subject to least-privilege access and auditing. The move reflects a broader industry shift toward identity for machines and agents as agentic AI moves into sensitive public-sector and healthcare workloads. By anchoring agent governance to existing compliance boundaries, Okta aims to lower the barrier for regulated organizations to adopt AI agents safely.
💡 Bringing agent identity lifecycle management inside FedRAMP/HIPAA boundaries lets security and platform teams govern non-human/agent identities with the same least-privilege, audit, and provisioning controls they already run for humans — removing a key compliance blocker to deploying agents in regulated environments.
This digest was collected from RSS feeds and summarized by AI (Claude). See the original links for full details.