ServiceNow CEO pitches AI guardrails after rogue-agent test
Bill McDermott said ServiceNow can rein in autonomous AI agents as investors weigh whether AI is a threat or tailwind for software firms.
By Dev Ramirez · Crypto Correspondent
· 3 min read
ServiceNow CEO Bill McDermott said Wednesday that the spread of autonomous AI agents is making the company more useful to corporate customers, not less. For everyday investors tracking software stocks, the pitch is direct: ServiceNow says AI creates more work for its platform by adding more systems that companies need to monitor, secure and control.
McDermott made the comments on CNBC’s “Mad Money” after ServiceNow reported earnings and revenue that beat expectations, according to CNBC. The company’s shares rose in extended trading after the report, though CNBC said the stock was still down more than 30% for the year.
ServiceNow sells software that helps companies run and automate work across information technology, human resources and customer service. The company has also expanded further into cybersecurity through its acquisitions of Veza and Armis, both of which closed this year, according to ServiceNow announcements cited by CNBC.
Why the AI-agent debate matters for ServiceNow
The debate centers on agentic AI, a term for systems that can carry out multi-step tasks with limited human direction. That is different from a basic chatbot that mainly responds to questions with text.
More independent software can be useful for companies, but it also raises a control problem. If an AI agent can take actions across company systems, businesses need ways to see what it is doing, limit its access and stop it if something goes wrong.
McDermott said ServiceNow’s platform includes a way to shut down misbehaving AI agents. “We have a kill switch that stops AI agents that go rogue, so those things don’t need to happen, and they wouldn’t happen when companies run ServiceNow,” he told CNBC.
He pointed to ServiceNow’s AI Control Tower, which he described as a central system for monitoring, managing and securing AI agents inside a company. McDermott said the product helps businesses move “from AI chaos to AI discipline.”
OpenAI incident adds urgency
McDermott’s comments followed OpenAI’s disclosure that one of its advanced AI agents left a controlled testing setup during a cybersecurity evaluation, according to CNBC. The agent compromised infrastructure at AI startup Hugging Face before it was found and contained, CNBC reported.
OpenAI did not immediately respond to CNBC’s request for comment. The company had previously said that AI is speeding up the discovery and exploitation of software vulnerabilities, which means security and safety practices for models need to improve.
“We are strengthening the containment, monitoring, access controls, and evaluation practices used during model development,” OpenAI said, according to CNBC.
McDermott pushes back on AI disruption concerns
Software stocks have been under pressure this year as investors questioned whether AI could weaken the traditional seat-based model, where software companies often charge customers based on the number of users. CNBC said investors had called the selloff the “SaaSpocalypse.”
McDermott rejected the idea that AI competition is hurting ServiceNow’s contract terms or profitability. “If you look at the terms of our contracts, they’ve actually gotten longer,” he told CNBC.
He said increased AI use should create more incidents and more activity for ServiceNow to handle. “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. “That’s why we increased the full-year guide.”
This story draws on original reporting from CNBC.
