Cisco's Liz Centoni Tells Partners AI Services Are Their Next Profit Engine

Cisco IQ and the partner revenue case

According to Cisco, Cisco IQ brings its support and professional services capabilities into a single platform. The stated aim is to help partners identify risk, automate lifecycle management and create recurring revenue opportunities.

Centoni said branded services built on top of IQ account for 54 percent of revenue and 73 percent of margin for Cisco's top partners. She said IQ removes manual scoping, which lets partners scale those services more profitably.

She also described a path from insight to billable work. Automated assessments cover vulnerabilities in customer environments, last day of support, priority bugs, software conformance and operating system upgrade readiness. Cisco wants to turn that telemetry into funded modernization engagements, because, in her words, customers must modernize both to run AI workflows and to protect against AI-enabled attacks.

On the economics, Centoni pointed to deal registration and DSA eligibility, which she said keep margins protected for partners who identify and develop lifecycle opportunities. She argued that recurring contracted revenue creates an EBITDA multiplier, which one-off assessments do not.

Customer results cited by Cisco

Centoni gave two customer examples. One customer, which will be discussed at Partner Summit, said its security advisory response time fell from four days to 10 minutes across 11,000 devices. Another saw support tickets fall 10 percent and network reliability productivity rise 25 percent. These figures are Cisco's own and come from the interview, not from an independent evaluation.

Why AI pilots stall, according to Cisco

Asked why many enterprises remain stuck in AI pilot mode, Centoni said the failure comes from laying AI onto broken workflows. In her view, adding AI to a legacy process only makes a broken experience run faster. She said small pilots stall when companies treat AI as an add-on instead of redesigning the workflow.

She said the constraint is not the model or the infrastructure but the human feedback loop. Agents differ from conventional software, she said: they may launch at 60 percent of their capability and reach 90 percent-plus once they are usable at scale, which depends on users feeling safe to give honest feedback.

Cisco applied this internally. In its renewals organization, 40 percent of the team's time went to pulling data from 50 different sources and creating vanilla proposals. Centoni said Cisco questioned each step, decided where agents would do the task and where the human boundary lay, and let AI surface context so people could focus on judgment, relationships and customization. She said Cisco's CX organization, which has more than 20,000 people, is more than a couple of years into this work.

The "Switzerland of AI" and an AI-native services model

Asked about Cisco's positioning as the "Switzerland of AI", Centoni described four principles for moving its services organization to an AI-native model:

• AI changes the cost structure of expertise, so workflows limited by human availability should be redesigned.
• The status quo is not an option.
• Today's quality is the floor, with AI raising the ceiling.
• The type of talent matters as much as the technology.

She said 95 percent of Cisco's support work is now AI-enabled. According to her, an AI assistant pulls in customer context so that the first 45 minutes of data gathering shrinks to a couple of minutes, and it suggests next best actions to engineers based on more than 40 years of experience.

What it means for partners

The message to the channel is that AI-driven services, built on telemetry and tied to incentives, can shift partner business from one-off assessments to continuing contracts. Cisco says IQ is generally available now, and Partner Summit 2026 is where the company is expected to expand on these themes. The interview excerpt provided does not cover pricing or specific program terms.