Workday and ServiceNow Launch Agent Platforms to Control What AI Agents Can Do

Companies in this story: Workday, ServiceNow, Nvidia

Two of the biggest names in enterprise software spent the same week answering the question every IT leader is now asking about AI agents: how do we stay in control?

Workday: autopilot agents with guardrails

At its Rising conference on October 13, Workday launched autopilot agents that work in the background across hiring, payroll, accounting and IT, following company policies, permissions and approvals. Its Sana experience becomes the new Workday home screen, bringing tasks, decisions and agent activity into one place.

Workday also said its Trusted Agent Platform, where organisations build and run agents governed by their business rules, is now "headless", so teams can use their own tools and models. Availability is staged: the Talent Acquisition Agent's autopilot is available to early adopters now, with payroll, accounting and deployment autopilots following for early adopters in early 2027.

ServiceNow: a passport for every agent

On October 15, ServiceNow unveiled AI Passport, a signed credential that defines an agent's identity, what it is allowed to do, which systems it can access and who is accountable for it. Passports are issued through its AI Control Tower and checked as work happens, including for agents from other vendors.

In one demo reported by SiliconANGLE, a customer service agent automatically approved an $18 refund while a $40 request was routed to a human, because it had been configured to approve refunds only below $25. ServiceNow said AI Passport, shadow AI detection and AI cost management are expected in the fourth quarter of 2026, with Nvidia as a design and launch partner.

Why control is the new battleground

Early AI agents were judged on what they could do. Enterprise buyers are now judging them on what they cannot do without permission. As agents start approving refunds, changing records and moving money, companies need clear answers to three questions:

1. Identity: which agent did this, and who owns it?
2. Permissions: what is it allowed to do, and where are the limits?
3. Audit: can we see and explain every action afterwards?

Both announcements are direct answers to those questions.

What it means for SaaS companies

• Governance is now a sales requirement. If your product includes agents, expect enterprise buyers to ask about permissions, approvals, audit logs and human handoff.
• Integrations matter. Platforms like ServiceNow are positioning themselves to govern everyone else's agents. Being compatible with those control layers may become a deciding factor.
• Trust content wins deals. Clear, specific security and AI-governance pages help both human buyers and AI assistants recommend you.

For a wider view of the tools in this space, see our upcoming guide to AI agent security and governance platforms, and explore how we help SaaS teams build content that earns trust.