Cisco's Winning AI Formula: Real CX Problems, Practical Solutions

Forbes

16-06-2025

Companies with the best customer experience focus on consistency, clarity, and a mindset of continual improvement. Most enterprise AI initiatives fail not because the technology doesn't work but because companies chase broad or ill-defined use cases instead of addressing a real problem. For example, many organizations have built chatbots that wow in demos but frustrate users in practice.
Liz Centoni, Cisco's Chief Customer Experience Officer
When Liz Centoni, Cisco's Chief Customer Experience Officer, talks about solving "boring problems," she's not being modest—she's highlighting a fundamental truth about artificial intelligence that most companies miss. While the tech world obsesses over flashy AI demos and theoretical capabilities, Cisco quietly built practical and measurable AI use cases that make it easier for its enterprise customers to use and troubleshoot their Cisco environments.
"We're solving the most boring problems that are instrumental to our customers' operational environments—problems everybody's been circling around for years," Centoni explained during an industry analyst breakout at the Cisco Live conference in San Diego. What are examples of these "boring" problems? Configuration errors that cause 25% of all support cases. Network professionals spending up to 50% of their time on manual tasks and minimizing security breaches caused by human error.
The results speak volumes: Cisco has achieved a 22-25% decrease in low-severity support cases and a 10% reduction in high-severity cases year-over-year. Additionally, its AI-powered renewal process has reduced the time its customer success teams spend on data gathering from 40% to under 5%, freeing them to focus on actual customer relationships.
By addressing the low-hanging fruit of basic support issues, Cisco can focus its support teams' time on more complex problems while also enhancing its sales process.
During her Cisco Live keynote, Centoni shared that Cisco's customer experience strategy centers on three core areas that any company can adapt to its customer experience challenges:
The most tangible impact comes from what Cisco calls "services as code"—integrating AI-powered testing into deployment pipelines to catch configuration errors before they cause outages. "We can envision a future where we go from configuration chaos to configuration confidence," Centoni explained.
This isn't just about finding defects. The system proactively validates configurations against established best practices and operational requirements specific to each customer's environment. One customer who adopted this approach summarized the value: "Security, resiliency, consistency—you delivered all three."
The broader lesson: AI's value often lies not in replacing human decision-making but in preventing the human errors that cause the most expensive problems.
Cisco recognized that customers were drowning in multiple interfaces and disconnected tools. Like many large technology vendors, Cisco aims to simplify the customer experience (CX) by offering a unified, AI-powered interface that provides a "hyper-personalized view into your entire Cisco environment," as Centoni described it.
This interface doesn't just aggregate information—it understands context. It can identify which devices are approaching end-of-support, suggest remediation scripts for security vulnerabilities, and even generate compliance reports tailored to specific regulatory requirements.
The key insight: AI's real power in simplification comes not from hiding complexity but from making complex information actionable and relevant to each user's specific context.
Personalization isn't a new concept, but it's proven elusive in both consumer and B2B sales. Cisco has created what they call an "adoption agent" that digitalizes customer intent and creates personalized onboarding journeys. Rather than providing a standard set of features, the system aligns adoption with each customer's specific goals and key performance indicators (KPIs).
"We're digitizing the customer's intent, the KPIs, the outcomes, and then we're helping them adopt the features that tie up to that intent, not just a whole standard set of features," Centoni explained.
Breaking down data siloes was a key theme of most technology vendor's presentations in this spring's technology conference circuit. Cisco also showcased how AI could help the company connect and analyze data across various sources. This strategy represents a shift from product-centric to outcome-centric customer success enabled by AI's ability to process and connect disparate data sources.
In 2025, a technology conference can't be complete without sharing a vision for Agentic AI. Cisco was no exception. While there is still some debate over the definitions of Agentic AI, most technology companies define it as a system of AI agents 'designed to act autonomously, making decisions and taking actions to achieve goals with limited human oversight. Unlike generative AI, which focuses on creating content, agentic AI focuses on doing by executing tasks and solving problems. It perceives its environment, reasons about it, and acts upon it, often without direct human intervention." Agentic AI concept is both empowering and terrifying to organizations that want to reap the productivity of agents but need to minimize the risk of fully autonomous workflows.
During the analyst conference at Cisco Live, Centoni shared a balanced approach to moving into Agentic AI. She said, "We want our teams to think about it as augmentation," Centoni emphasized. "I would love to be in a space where instead of asking for an intern to help them do their job, everyone in my team could spin up an agent to be able to help them with tasks."
Agentic AI agents can operate like capable colleagues, understanding their context, making informed decisions, and coordinating multiple tasks to achieve a goal. Carlos Pereira, Cisco's Fellow and Chief Architect for Customer Experience, explained the distinction: "The way we look at it is the way we have been using traditional AI as a tool. The way we expect to use agentic AI is where it becomes a teammate."
This shift from tool to teammate enables what the technology industry refers to as "ambient agents"—AI systems that operate in the background, triggered by events rather than direct commands. As Harrison Chase, CEO of LangChain (a key partner for Cisco in building these systems), described during the Cisco Live keynote: "We define ambient agents as agents that are triggered by events, run in the background, but they're not completely autonomous."
The power of this approach becomes clear in practice. Instead of a customer reporting a network issue and waiting for a human to diagnose it, Cisco's ambient agents can detect the problem in real-time, analyze historical data and best practices, and provide personalized recommendations—all before the customer even knows there's an issue.
While Cisco's efficiency gains are impressive, the real return on investment extends beyond traditional metrics. Centoni noted that customer satisfaction consistently improves when solutions are found through AI-enabled methods. AI will also change the nature of work itself at Cisco. "Reducing cognitive load and friction enables my teams to get more creative in how we solve our customers' problems," Centoni observed. "They can use that (extra) time for learning. They can use that time to balance work and life."
Agentic AI offers significant potential business impact, where AI not only enhances existing processes but also enables entirely new ways of creating value. When routine tasks are automated, human workers can focus on the complex, creative problem-solving that drives real competitive advantage.
Cisco also offers practical lessons for any organization looking to transform customer experience with AI:
Rather than asking, "What can AI do for us?" Cisco asked, "What problems do our customers and employees face every day?" This question led Cisco to focus on configuration errors, manual tasks, and data silos—initial use cases that may seem unglamorous but can deliver high impact rapidly.
"We are thinking about autonomous in terms of tasks that augment what our teams do," Centoni emphasized. This approach reduces resistance, maintains quality control, and often delivers better results than fully automated systems. Over time, there will be opportunities to have more autonomous systems; however, Cisco's strategy offers a more pragmatic approach to minimizing risk today.
Unlike traditional software that follows a "build it, ship it, maintain it" cycle, AI systems require continuous improvement. "It's build it, improve the accuracy of it... it's continuous learning," Centoni noted. Businesses need to design processes for ongoing feedback and refinement.
With customer relationships at stake, Cisco maintains human oversight at critical decision points. "The decision is never up to the agent, per se. The decision is up to the human at the end of the day," Centoni explained. The balance between AI capability and human control builds trust with both employees and customers.
While cost savings matter, the real value lies in enabling new capabilities. Cisco's agents don't just handle support cases faster—they can predict and prevent issues that would never have been caught manually.
Cisco's vision extends beyond current capabilities to what Centoni calls "intelligent anticipation"—systems that understand customer environments so deeply that they can resolve problems before customers are even aware of them.
"Our goal, whether it's a customer who spends a few thousand dollars or a customer who spends a few billion dollars with us: we want them to feel like they're our only customer because we know their environment. We know them so well, sometimes even better than they do themselves," Centoni explained.
The vision of hyper-personalized, predictive customer experience represents the true promise of AI in business—not replacing human relationships but making them more meaningful by removing friction and adding intelligence to every interaction.