From Prompting to AI Workflow Execution in Customer Service
AI customer service agents are software agents that use artificial intelligence to execute multi-step customer experience workflows across channels, replacing one-off prompts with repeatable automations that resolve inquiries, trigger back-office actions, and feed analytics without needing constant human instruction. This shift is reshaping how contact centers think about value. Instead of paying for static software seats, enterprises are starting to fund AI workflow execution that ties directly to resolved cases, campaign throughput, or operational savings. Outcome-based AI pricing is emerging as a natural fit: buyers do not only want faster replies, they want measurable outputs like more concluded experiments, higher campaign production, or reduced after-call work. As unified communications and contact center stacks converge, platforms that can bundle AI customer service agents, telephony, routing, and customer experience automation in one CCaaS platform revenue model are gaining an edge over fragmented tools.
Zoom’s ZCX Push: CCaaS Revenue Built on Paid AI
Zoom’s latest quarter shows how AI can turn contact centers into a revenue engine rather than a line item. The company reported Q1 revenue of USD 1.24 billion (approx. RM5.73 billion), up 5.5% year over year, with enterprise revenue growing 7.2% and making up 61% of the total. Eight of the top ten Zoom Contact Center (ZCX) deals displaced incumbent CCaaS providers, and paid AI was present in nine of those top ten ZCX wins, signalling that AI is now a funded core capability. According to Zoom, paid monthly active users of AI Companion surged 184% year over year as customers expand AI-powered customer interactions. Four of the top ten ZCX deals also included Zoom Phone, and four of the top ten Zoom Phone deals included ZCX, a pattern that shows buyers want unified UC and CC stacks where AI agents, routing, and agent workflows operate in one environment.
Optimizely Opal: AI Agents as a Marketing Workflow Engine
Optimizely’s Opal AI agent platform shows what happens when teams move beyond “AI helps me write faster” to full workflow execution. The company reports that annual recurring revenue for Opal grew 42% quarter over quarter as customers ran more work through agents embedded in core digital experience workflows. Nearly 1,700 customers have built more than 4,000 custom AI agents and executed over 172,000 runs across experimentation, content, campaigns, and reporting, with more than 97% of activity driven by customer-built agents rather than prebuilt assistants. Around 32% of executions involve multi-step tasks, which shows that AI workflow execution is handling end-to-end work, not single prompts. Downstream performance indicators echo this: concluded experiments are up 38.0%, campaign production jumps 85% when Opal is paired with the CMP, and digital asset reuse rises 57%, linking AI customer service agents and marketing agents directly to measurable capacity gains.
Outcome-Based AI Pricing and the CCaaS Platform Battle
Both Zoom and Optimizely point toward the same business model shift: outcome-based AI pricing anchored in workflow execution rather than seat licences. Zoom is openly talking about monetizing AI on outcomes, not seats, using AI Companion and ZCX to move from call transcription into automated post-call work, CRM updates, and ticket routing. Optimizely positions Opal as an orchestration layer that moves work reliably across CMS, DAM, analytics, and experimentation tools, where budgets are easier to defend because throughput and consistency can be measured. This puts pressure on legacy CCaaS providers that still sell disconnected telephony and basic bots. As unified platforms bundle AI customer service agents, routing, analytics, and customer experience automation into one CCaaS platform revenue story, buyers gain a clearer line from AI spend to experiments concluded, campaigns shipped, or cases resolved, which strengthens the case for larger, multi-product platform deals.