Agentic AI Contact Centers: From Scripts to Self‑Directed Work
Agentic AI contact centers use autonomous contact center agents that can interpret intent, take multi-step actions across systems, and decide when to involve humans, shifting automation from rigid scripts and menus toward self-directed resolution and intelligent AI agent handoffs that are measured by outcomes instead of deployments alone. Agentic AI is pushing contact centers past scripted chatbots toward autonomous problem-solving — but operationalizing that shift remains uneven. The defining challenge for CX leaders in late 2026 is closing the distance between technology investment and measurable outcomes, as governance, handoff orchestration and workflow integration still slow down ROI for many organizations. At the same time, agents now handle up to 40% of inquiries across multiple channels, and projections suggest that could reach 80% by 2029. The question is which contact center AI platforms are setting themselves up to win that future.
Zoom: Prompt-Built Agents and Pay-per-Resolution Bets
Zoom is betting that agentic AI will only scale if non-technical CX teams can build and tune agents as easily as writing an email. Zoom Communications on June 22 announced new AI capabilities for Zoom Virtual Agent (ZVA), including Agent Architect, which generates production-ready voice or digital agents from simple prompts. Rather than celebrate deployments, Zoom is tying success and billing to resolved or successfully routed interactions through outcome-based pricing for ZVA. That is a bold signal: if the agent does not solve problems, Zoom does not get paid as much. An Agent Performance Suite that simulates scenarios, validates results and provides real-time dashboards pushes customers toward continuous optimization instead of one-off projects. With AI Companion paid users up 184% year over year on Q1 FY2027 revenue of USD 1.239 billion (approx. RM5.69 billion), Zoom has both pressure and runway to prove that pay-per-resolution can be a sustainable model.
RingCentral and Five9: Intelligent IVR Replacement or New Sprawl?
RingCentral and Five9 are taking the most aggressive swing at legacy IVR, pitching autonomous contact center agents as the new front door. RingCentral has repositioned as an AI-native communications platform, and on June 23 it expanded AIR Pro to add native AI agents, autonomous outreach, intelligent handoffs and new analytics to RingCX. These agents handle inbound and outbound interactions across voice and digital channels and aim for end-to-end resolution while preserving context when conversations transfer to live agents. New capabilities are in beta, with general availability planned for the second half of 2026. Five9’s new Voice AI Agents target enterprises looking to move beyond scripted bots and legacy Interactive Voice Response (IVR) systems with agentic AI for customer self-service. Multi-agent orchestration lets coordinated AI agents work across complex, multi-step customer journeys while providing context-rich warm handoffs so AI and human agents collaborate on a unified platform. The risk for buyers is swapping one generation of fragmented IVR trees for a new generation of fragmented AI flows if governance and design discipline lag the technology.
Verint and CallMiner: Orchestrating Work vs. Augmenting Humans
Verint and CallMiner are pushing against full autonomy in different ways: one by obsessing over the work, the other over the worker. Verint on June 23 unveiled Workforce Intelligence, Desktop Intelligence and Quality Intelligence at its Engage 2026 conference, all built on Verint Agent Factory. Agent Factory lets enterprises build, configure and manage blended human and AI workforces within its CX Automation Platform, routing work to human staff when needed. Workforce Intelligence controls staffing, Desktop Intelligence captures desktop activity as AI-ready data, and Quality Intelligence connects agent speech with system actions to flag outcome gaps. This unifies behavior, processes and outcomes rather than treating AI agents as a separate black box. CallMiner, by contrast, expanded its RealTime platform with agentic AI that gives contact center agents on-demand, context-aware guidance during live interactions. Building on agent-initiated AI guidance, the company doubles down on human-in-the-loop oversight to maintain transparency, explainability and compliance. In effect, Verint orchestrates work across humans and AI; CallMiner empowers humans and keeps AI on a short leash.
Outcome-Driven Agentic AI: Who Will Own the Resolution Metric?
The most important shift in agentic AI contact centers is commercial, not technical: vendors are starting to sell outcomes, not seats. Zoom’s outcome-based pricing for ZVA, which ties billing to resolved or successfully routed interactions, is the clearest sign that pay-per-resolution models are emerging. If that spreads, the industry will finally be forced to measure containment, successful AI agent handoffs and customer effort with the same rigor as talk time. Agentic AI orchestration is already moving from pilots to production and generating measurable gains in productivity, cost savings and compliance. Five9 expects its Voice AI Agents to handle more than 100,000 service calls by year-end while improving the customer experience. Yet the race is far from settled. RingCentral and Five9 are vying to own intelligent IVR replacement, Verint wants to be the operating system for all work, and CallMiner is staking out the explainable, human-centric flank. Buyers should judge them on a simple question: who is willing to share the risk by getting paid only when the AI resolves real problems?
