AI agents are delivering dramatic productivity gains for some businesses, but weak governance risks preventing companies from deploying them, according to ServiceNow executive Paul Fipps.
Fipps, president of global customer operations at ServiceNow, said Standard Chartered Bank was achieving a 77% “deflection rate” across employee requests. The figure refers to requests handled without human intervention.
Speaking at the 2026 AIQ Summit, Fipps said: “I always say, if you get 25% to 30% deflection in those use cases, it’s game-changing. If you get 75% to 80%, it is life-changing.”
He said the most valuable applications of AI were often “horizontal”, meaning they could be used across several departments rather than being confined to a single business function.
That can include automating requests involving human resources, legal services and technology. Fipps said employee experience was not limited to HR, arguing that broader use of AI could deliver “tremendous ROI”.
Businesses are redesigning processes around AI agents
Fipps said companies would gain more by rethinking how work was organised than by simply adding AI to existing systems.
“If I just automate or drive autonomous agents to drive the current processes I have, that’s a choice that might get some productivity bump,” he said. “But the ones that are actually rethinking and reimagining the business processes, and then leveraging AI inside those processes—whether they are vertical or horizontal—I think are really making the difference.”
He said businesses needed to consider both how to use the technology efficiently and how employees could be redeployed to carry out “new and innovative things”. Productivity, he added, should not be treated as the sole objective.
Companies also need to decide how to use the time and resources released by AI, with the strongest results coming when businesses begin with customer needs and work backwards to identify where technology can have the greatest effect.
ServiceNow announced earlier this year that its entire product portfolio would be AI-enabled, describing the company as an “AI control tower for business reinvention”.
AI governance is becoming a barrier to deployment
Fipps warned that governance and trust were becoming central to the adoption of AI agents, particularly in regulated industries. Poorly controlled systems could breach compliance rules and damage customer confidence, he said.
He recalled speaking to the chief technology officer of a large bank in India that had built about 40 custom agents but had not deployed any of them.
“I haven’t released any of them because I can’t rely on humans to control these agents, and I’m a regulated bank in India. And so I actually have these 40 fantastic custom-built agents that I can’t release,” the CTO told him, according to Fipps.
“I see governance and trust driving speed,” Fipps said, arguing that effective oversight could allow businesses to deploy AI more quickly rather than holding it back.
The number of active enterprise agents is estimated to rise from 28.6 million in 2025 to 2.2 billion by 2030, according to the International Data Corporation. Fipps said the scale of that expansion would make maintaining human control an increasingly significant challenge.
