From Deflection Rates to CX Business Outcomes
For years, AI in customer service was judged by one basic measure: how many inquiries it could deflect from human agents. Kustomer’s new Architect platform argues that this support deflection metric is not only incomplete, it can be actively misleading. Deflection simply reports how many conversations never reached a human; it says nothing about whether customers were satisfied, stayed loyal, or spent more. A bot that blocks access to agents might reduce ticket volume while quietly driving churn. Architect instead places agentic AI customer service inside a unified stack of data, workflows, knowledge, and human agents, and then ties activity to outcomes like retention, loyalty, efficiency, and revenue. The message to CX and finance leaders is blunt: stop optimizing for a cost line and start measuring AI customer experience automation by the business value it creates across the customer lifecycle.
Nvidia’s 66% Support Deflection Shows Agentic AI at Enterprise Scale
Nvidia’s latest earnings call framed agentic AI as an operational necessity rather than a lab experiment, and customer support is a prime example. The company highlighted a 66% support deflection rate achieved through agentic AI, using the same automation playbook internally that it promotes to other enterprises. This marks a shift from static chatbots to AI systems that execute work: routing issues, resolving known problems, and assisting employees. Leadership described demand for such automation as “parabolic,” arguing that tokens are now profitable because AI agents can perform productive tasks, not just generate text. For CX leaders, this signals mainstream acceptance that AI agent workflows belong on the operating expense line as a lever for productivity, not just a pilot project. The implication is clear: board-level discussions about support deflection metrics are expanding to include how AI reshapes staffing, tooling, and broader experience strategies.
Optimizely’s Opal Shows AI Agents Becoming Workflow Engines
Optimizely’s Opal platform illustrates how AI agents are moving beyond response automation into full workflow execution. The company reports 42% quarter-over-quarter growth in annual recurring revenue for Opal, powered by teams building more than 4,000 custom AI agents and running over 172,000 executions. Crucially, more than 97% of activity comes from customer-built agents, showing that organizations are investing in reusable AI agent workflows tailored to their own stacks. Around 32% of executions involve multi-step tasks, such as orchestrating experiments, campaigns, and reporting end-to-end rather than answering a single prompt. This shift correlates with higher throughput metrics: more concluded experiments, higher campaign production when paired with Optimizely’s content tools, and significant gains in digital asset reuse. Together, these trends show CX and marketing leaders using agentic AI not just to respond faster, but to redesign how work flows across their digital experience platforms.
From Cost Center to Profit Engine: How Support Models Are Being Rewritten
As agentic AI becomes embedded in platforms from Kustomer to Optimizely, it is also reshaping how companies like Text (parent of LiveChat, ChatBot, and HelpDesk) and Zoom position customer service. Instead of treating support as a pure cost center to be minimized through aggressive deflection, these firms are emphasizing CX business outcomes such as expansion revenue, improved retention, and higher lifetime value. AI customer experience automation can proactively surface upsell opportunities, accelerate resolution for high-value accounts, and feed insights back into product and sales. That changes staffing assumptions: human agents are reserved for complex, high-impact work, while AI handles scalable, repeatable tasks. Evaluation criteria follow suit. Rather than asking, “How many tickets did we avoid?” leaders are asking, “Which AI agents drive loyalty, renewals, and growth?” In this model, support deflection metrics are just one input into a broader profit-focused CX strategy.