ServiceNow’s $1 billion AI ACV shows demand, not yet governance control

The $1 billion AI figure should therefore be read as annual contract value—a contracted annualized measure—not as a separately reported quarterly revenue stream. ServiceNow said the number of customers running its agentic AI offerings in production increased ninefold over nine months, but did not provide the starting or ending customer count in the retained reporting.

Company-reported AI Control Tower interface supplied by ServiceNow in its May 2026 announcement. Source: ServiceNow Newsroom.

Control Tower is the strategic claim

Amit Zavery, ServiceNow’s president, chief product officer and chief operating officer, leads the platform, product, engineering and cloud-infrastructure organization. He joined in October 2024 after nearly six years leading platform work at Google Cloud and almost 25 years at Oracle, where he became executive vice president of product development, according to ServiceNow’s appointment announcement. That background explains why the current argument is about an enterprise platform layer rather than a single model or chatbot.

Zavery told an interview that customers combining technologies in-house can encounter visibility, security and governance problems. He also made the important limiting point: existing ServiceNow customers remain the biggest driver of AI growth. That makes the installed base an advantage, but it leaves the company’s reporting unable to distinguish a new governance-led sale from an expansion of workflows a customer already runs on the platform.

ServiceNow says its May AI Control Tower expansion added 30 enterprise integrations spanning AWS, Google Cloud and Microsoft Azure, as well as applications including SAP, Oracle and Workday. The company says the product, first introduced at Knowledge 2025, can cover assets beyond ServiceNow, nonhuman identities and connected devices. In its announcement, it also said the enhancements were entering Innovation Lab in May and were expected to become generally available in August.

Those are product and availability claims from ServiceNow, not independent evidence that the system provides equivalent governance across every external model, agent and infrastructure environment. The same announcement names both a claimed advantage over standalone governance tools and risks that future capabilities, sales and regulatory conditions could differ from expectations. The competitive comparison is therefore still the company’s own position, rather than a demonstrated market result.

What would turn the pitch into proof

The next decision for buyers is whether Control Tower reduces the practical burden of managing mixed AI estates enough to justify becoming a platform commitment, rather than another feature inside an existing ServiceNow deployment.

Evidence that would clarify that decision is straightforward: independently documented use across non-ServiceNow models and agents; customer retention and expansion tied specifically to Control Tower; and a disclosure separating AI annual contract value won from new customers from value expanded within the installed base. Until then, the quarter supports a conclusion about demand for ServiceNow’s AI offerings, but not a conclusion that it has established governance control across enterprise AI.