The CX market is experiencing a decisive shift toward agentic AI as a core operational lever rather than an experimental feature. Zoom's positioning of ZCX as an AI-first replacement for legacy contact center stacks—with eight of its top ten deals displacing incumbents and paid AI embedded in nine of those same deals—signals that switching costs are eroding faster than previously assumed. More tellingly, Salesforce's Agentforce has crossed $1 billion ARR whilst handling double the inquiry volume of human agents across its own support operations, with customers like Florida Prepaid now running 75% of business-hour calls autonomously. This convergence matters because it reframes the competitive landscape: teams evaluating contact center platforms are no longer comparing feature parity or integration depth, but rather the maturity and ROI of autonomous agent capabilities. For organisations already running Agentforce or similar agentic platforms, the question becomes whether competitive pressure will force rapid feature adoption across the rest of the stack, or whether best-of-breed point solutions will fragment the market further.
Simultaneously, Sprinklr's acquisition of ViralMoment exposes a critical blind spot in how most CX teams listen to customers. As social engagement migrates toward visual-first channels—TikTok, Instagram Reels, YouTube Shorts—traditional Voice of the Customer programmes remain anchored to text-based feedback and written reviews. This gap is not merely a data collection problem; it represents a fundamental misalignment between where customers express sentiment and where enterprises are listening. Sprinklr's move to ingest multimodal intelligence across video, images, audio, and text acknowledges that customer intent and frustration are increasingly non-textual, yet most incumbent platforms lack native capabilities to process these signals at scale.
The final piece of this week's narrative—IBM and Red Hat's $5 billion Project Lightwell initiative—introduces an underappreciated risk dimension. As frontier AI systems like Anthropic's Claude Mythos accelerate vulnerability discovery in open-source software, CX teams relying on cloud-native stacks face a new class of security exposure. The concern is not hypothetical: Mythos has already identified nearly 3,900 high- or critical-severity vulnerabilities in open-source libraries. For CX leaders, this means security and compliance considerations are no longer separable from AI adoption decisions. Teams deploying agentic AI must now evaluate not only agent performance and cost per interaction, but also the vulnerability surface of the underlying infrastructure—a complexity that favours vendors with deep security investment and transparency around their dependency chains.
From Zoom’s push to position itself as a revenue engine to a major Sprinklr acquisition, here are extracts from some of this week’s most popular news stories. Zoom Is Turning CX Into A Revenue Engine, Should Legacy CCaaS Be Worried Zoom’s Q1 FY2027 earnings call made one thing uncomfortably cl