AI agent identity: ServiceNow adds verifiable AI Passport to control tower

What happened

ServiceNow has added governance features to its AI Control Tower and renamed the product Super Intelligence Control Tower (SICT). The headline addition is AI Passport, a new component that gives every AI agent a verifiable identity and provides runtime control over its activity, whether the agent was built by ServiceNow or a third party.

According to the reporting, an AI Passport is a cryptographically signed credential that is checked at runtime to confirm the scope the enterprise has authorised. An agent that stays within its passport scope can operate freely. If it attempts to deviate, defined actions apply.

ServiceNow is also updating EmployeeWorks to make it easier to build AI agents. Computerworld and InfoWorld both covered the announcement, which the company framed around the shift from agent pilots to production deployments.

Why it matters

The move reflects a change in what enterprises need from AI agents. Once agents are running autonomously across business systems, the central question is no longer only whether an agent can complete a task. Leaders also need to know which agents are active, what they are doing and which systems they can reach. Nenshad Bardoliwalla, SVP of AI product management at ServiceNow, put the risk directly:

If companies can't identify which agents are running, what they are doing, and what systems they can access, then scaling quickly becomes a risk.

Treating agent identity as infrastructure, rather than a policy document, changes how organisations design and operate AI-enabled services. Agents that handle requests, records or workflows touching customers and staff need the same traceability and clearly bounded permissions that organisations already expect of people. Runtime checks of this kind also make it possible to govern agents as they act, not only after an incident.

The Renascence take

Most coverage of this announcement will focus on the product features. The more useful question for service leaders is what happens to trust when a customer-facing process depends on software that nobody can fully see.

Verifiable identity is a foundation, not a finish line, because an authenticated agent can still make a poor decision inside its permitted scope. The behavioural principle underneath is that people trust systems they can understand and predict, and that predictability has to be designed in, not assumed. A customer-obsessed operator should map every agent to a named owner, a documented scope and a clear escalation path before extending its reach. Then test how the agent behaves at the edges of that scope, where experience failures tend to surface first.