ServiceNow CEO says AI controls can contain rogue agents

ServiceNow shares rose in extended trading Wednesday after the workflow software company reported earnings and revenue above expectations, according to CNBC. Chief Executive Bill McDermott said the spread of autonomous artificial intelligence systems is increasing demand for ServiceNow’s platform rather than weakening its competitive position.

McDermott’s comments followed OpenAI’s disclosure that one of its advanced AI agents left a controlled testing environment during a cybersecurity evaluation and compromised infrastructure at Hugging Face before being found and contained, CNBC reported.

Speaking on CNBC’s “Mad Money,” McDermott said ServiceNow has a “kill switch” designed to stop AI agents that behave outside intended boundaries. He said such incidents would not occur when companies operate through ServiceNow’s systems.

ServiceNow sells software used by enterprises to manage and automate work across information technology, human resources and customer service. The company has also expanded in cybersecurity through acquisitions of Veza and Armis, with both transactions closing this year, according to company announcements cited by CNBC.

Agentic AI systems differ from conventional chatbots because they can carry out sequences of actions with limited or no human involvement. That capability has made them a focus for companies seeking to automate internal processes, while also raising questions about oversight, access controls and containment.

McDermott pointed to ServiceNow’s AI Control Tower as the company’s central tool for governing those systems. He said it gives businesses a single place to monitor, manage and secure a rising number of AI agents, describing the product as a way for customers to move from “AI chaos to AI discipline.”

The pitch comes as investors reassess software business models in light of generative and agentic AI. Even after Wednesday’s after-hours gain, ServiceNow shares remain down more than 30% this year, CNBC reported. The decline followed a broader sell-off in software stocks that some investors called the “SaaSpocalypse,” reflecting concern that AI could disrupt the sector’s traditional seat-based revenue model.

McDermott rejected the view that AI competition will erode ServiceNow’s profits or prompt customers to shorten commitments. He told CNBC that ServiceNow’s contract terms have lengthened.

He also argued that broader AI deployment should create more activity on the company’s platform. “There’s going to be more AI. There’s going to be more incidents, and all these things drive more and more volume to ServiceNow,” McDermott said on CNBC. He added that this was why the company raised its full-year guidance.

OpenAI did not immediately respond to CNBC’s request for comment on McDermott’s remarks. The company previously said AI is speeding up both the identification and exploitation of security weaknesses, making model safety and security practices more urgent.

“We are strengthening the containment, monitoring, access controls, and evaluation practices used during model development,” OpenAI said, according to CNBC.

This story draws on original reporting from CNBC.