New Cisco Research Reveals Overwhelming Demand for Agentic AI to Transform the Way Technology Vendors Deliver Customer Experience
With agentic AI reaching a new level of maturity, we’re closer than ever to solving some of the most persistent customer pain points in enterprise environments. For instance, a significant share of network issues stem from misconfiguration, something agentic systems are poised to eliminate. That shift will lead to smarter networks, stronger security, and more productive teams. As an industry, we’ve been talking about these concepts for decades, and while we’ve made incremental progress, AI – and especially agentic AI – can make that vision a reality.
Imagine a world where customers never have to recap their situation to a new support agent, where issues are resolved before they are ever experienced by an end-user, and where every interaction feels tailored to individual needs and business outcomes. That’s not just a better experience; it’s a transformational one. This is the agentic future of Customer Experience (CX).
The benefits are clear and compelling, but how ready are organizations to adopt agentic AI? While customers are eager to reap the benefits of AI, how comfortable do they really feel interacting with AI agents instead of humans, and what are the effects on the customer relationship if something doesn’t work exactly as planned?
These are the types of questions we had in mind when we launched our most recent research study, surveying nearly 8,000 business and technical decision-makers worldwide on the future of agentic AI.
The findings confirmed some preconceived notions and totally challenged others. We know that AI is progressing incredibly fast and that it’s changing nearly everything about how we work, live, build, operate, and communicate. Those statements are practically clichés at this point but are still true. What we didn’t entirely anticipate is the ripple effect that AI is having across customer expectations, and the urgent need for technology vendors, many of whom are still only exploring the potential for agentic AI in customer experience, to accelerate their strategies.
Here are three key takeaways from the report.
Key Findings
- The Benefits of Agentic AI are Clear – and Demand is Soaring.
93% of global respondents believe that agentic AI will enable B2B technology vendors to deliver more personalized, proactive, and predictive services, improving productivity and creating time and cost savings. And the demand is intensifying. 56% of respondents expect their customer experience interactions will be handled by agentic AI within just 12 months, and that number jumps to 68% within three years. - Human Connection Remains Essential.
Despite the excitement around AI, people still need people. Customers want us to combine the efficiency of agentic AI with human connection to optimize customer experience. An overwhelming 96% of respondents said that human relationships are very important when interacting with B2B technology partners. This means that finding the right balance between human and machine is one of the most important challenges leaders face today. - Robust Governance is Non-Negotiable.
Customers understand, perhaps better than anyone, that AI must be deployed responsibly. A remarkable 99% of respondents stated that it’s important for technology partners to demonstrate robust governance arrangements to deliver ethical use of agentic AI. They demand transparency, security, and a commitment to eliminating data bias
Looking Ahead
At Cisco, we are drawing on four decades of data, our deep industry knowledge, and the trust of our customers to re-imagine Cisco Customer Experience as agentic AI-led. This research has validated our vision for the agentic future of customer experience but has also made one thing clear: we have to move faster.
Read the full research report
Agentic AI holds real promise, and the 68% projection reflects both innovation and the exciting potential to offload repetitive tasks while delivering 24/7 support. But alongside that optimism, there needs a clear view of the current realities. Independent reviews (example Yelp) show Cisco’s support services averaging just 3.1 out of 5 stars – couple this with studies showing 30–40% of customers report high satisfaction when interacting with AI-driven support (especially for complex or urgent issues) does not seem like a recipe for success.
Newer AI models, such as OpenAI’s o3 and o4-mini, exhibit higher hallucination rates—33% and 48% respectively—compared to older models like o1, which had a 16% rate. This increase in fabricated or inaccurate responses undermines trust at a time when confidence in AI systems needs to grow. These systems also depend heavily on historical support data, which means they often struggle with new products or edge cases. AI will undoubtedly transform support, but near-term success depends on strong human oversight, regular fine-tuning, and clear escalation paths.
Rather than replacing people, agentic AI should amplify the strengths of experienced support agents.