Artificial intelligence is rapidly moving from experimentation to production in enterprise contact centers. What began primarily as an opportunity to automate routine tasks and reduce labor costs is evolving into a broader transformation of customer engagement. AI can now analyze every interaction, assist human agents in real time, automate increasingly complex requests, and help organizations deliver more personalized experiences.
In a recent discussion with Zeus Kerrevala and Bob Laliberte at theCUBE Research Contact Center Summit, Five9 CEO Amit Mathradas outlined the company’s vision for this next generation of customer experience. Five9 calls its approach “Humantic”; a model in which human employees and agentic AI work together rather than compete for the same role.
The central message for enterprises is straightforward: AI can deliver measurable value in the contact center, but success will depend on selecting the right use cases, preparing the organization for change, and preserving access to human expertise when interactions involve complexity, value, or vulnerability.
Click below to watch the full video
Why Customer Experience Is an Early AI Use Case
Customer experience has emerged as one of the most compelling environments for enterprise AI adoption because contact centers combine high operating costs, large interaction volumes, and measurable performance indicators.
Mathradas noted that approximately nine out of every ten dollars of contact-center operating expense is associated with labor. That naturally makes the environment an attractive target for automation and optimization. However, he argued that the value proposition has already expanded beyond cost reduction.
“As AI has gotten better, you’re actually finding that it is not just removing costs, but it is revolutionizing how companies are leveraging contact centers and the experience that they are now providing,” Mathradas said.
Traditional quality-management processes might evaluate only a small percentage of recorded calls and use that sample to draw conclusions about the wider operation. AI enables organizations to analyze virtually every interaction. This can provide a more complete view of customer sentiment, agent performance, compliance, recurring problems, and opportunities for improvement.
AI can also improve routing by identifying the nature and context of a customer’s request and connecting that person with the agent best equipped to resolve it. As the cost per interaction declines, businesses may also be able to serve more customers across more channels and hours of the day.
Moving from Deterministic Bots to Agentic AI
The distinction between traditional automation and agentic AI is important. Earlier generations of chatbots and interactive voice-response systems were largely deterministic. They followed predefined scripts, decision trees, and “if-then” rules. These systems could handle simple, predictable requests but often struggled when customers expressed an issue in an unexpected way or needed help across multiple systems.
Agentic AI is more probabilistic and context aware. It can interpret intent, access approved organizational knowledge, consider possible courses of action, and determine whether to resolve the request, retrieve additional information, or transfer the interaction to a human.
This expands the range of interactions that businesses can automate, but it also increases the importance of governance. Organizations must establish what an AI agent is authorized to access, what actions it can take, when it must escalate, and how its performance will be monitored.
Improved models, lower latency, fewer hallucinations, and greater organizational maturity are helping enterprises become more comfortable with these deployments. Security, governance, and operational planning are increasingly being incorporated into projects from the beginning rather than treated as afterthoughts.
A Practical Path from Pilot to Production
For organizations beginning their contact-center AI journey, Mathradas advised against trying to transform the entire operation at once.
“Don’t boil the ocean. Don’t go say, ‘I’m solving everything with AI.’ Pick a use case, get it right, move to the next, move to the next,” he said.
That advice reflects a broader reality of enterprise AI adoption: clearly defined business problems typically produce better results than technology-led experimentation. Organizations should begin by identifying a measurable operational or customer challenge, such as reducing after-call work, improving agent consistency, increasing self-service containment, or shortening resolution times.
For businesses with limited AI experience, agent-assist capabilities may provide a logical starting point. These tools can recommend information, prompt agents during conversations, summarize interactions, and reduce administrative work while keeping a person in control. Agentic quality management offers another relatively manageable entry point because it can evaluate interactions at scale and identify coaching or process-improvement opportunities.
Voice AI agents potentially deliver greater efficiencies, but they are also more complicated to implement. They require careful design, integration, testing, tuning, and escalation procedures. Enterprises should therefore align the sophistication of the use case with their operational readiness.
Five9 is positioning its open architecture, professional services, and forward-deployed engineering capabilities as mechanisms to help customers build AI on top of their existing technology environments. This is significant because forcing an organization to replace its broader customer-experience stack can increase cost, extend implementation timelines, and delay measurable returns.
Measuring Outcomes Across Customers and Employees
Mathradas cited a large moving, logistics, and storage company as an example of phased adoption. The customer initially implemented AI agents and automated quality management before progressing to a generative AI voice bot. The deployment is expected to process approximately 100,000 calls annually and has achieved more than 50% containment for its targeted use case—reportedly exceeding its original goal.
The company also experienced improved customer satisfaction and lower agent churn. That combination is notable because it suggests AI can deliver benefits across multiple constituencies. Customers receive faster service, employees spend less time on repetitive work, and the business lowers the cost of handling routine interactions.
Enterprises should therefore avoid evaluating AI solely through labor reduction. A broader scorecard should include containment and resolution rates, customer satisfaction, employee retention, compliance, escalation quality, and the accuracy of the experience.
Preserving the Human Connection
The most important element of Five9’s Humantic vision may be its recognition that customers will not always want, or accept, an AI-only experience.
Mathradas identified complexity, value, and vulnerability as three conditions in which human involvement remains essential. A sensitive healthcare or insurance matter, a high-value customer relationship, or a complicated financial decision requires judgment, empathy, and accountability that cannot simply be optimized away.
“There is a misconception that all humans are going away in the contact center. That is not true,” Mathradas said.
The future contact center will likely use AI to handle high-volume requests, provide continuous availability, and augment employees. Human agents will increasingly concentrate on emotionally sensitive, financially important, or highly complex interactions. The critical architectural requirement will be a seamless transition between the two, with context preserved throughout the customer journey.
Why It Matters
AI is turning the contact center from a reactive service operation into a richer source of customer intelligence. Analyzing interactions at scale can help organizations capture sentiment, preserve conversational context, and orchestrate actions across CRM, commerce, support, and other systems of record.
The winners will not necessarily be the organizations that automate the most interactions. They will be those that apply AI selectively, measure outcomes rigorously, integrate it with existing workflows, and make it easy for customers to reach a capable human when circumstances demand it.
That balanced approach offers the most credible path to lowering costs without diminishing the customer experience, and to turning contact-center AI into a sustainable source of business value.
For more information on Five9 please visit their website.
