June 12, 2026
Your Customer Experience Roundup: Verizon Says the Quiet Part Out Loud, The Pricing Model Is Changing, and Your AI Agent Now Has an Agent
This week in customer experience, a telecom CEO said what most of his industry peers are thinking but declining to put on record; the earnings season continued its quiet argument for outcome-based AI pricing, and the contact centre software market grew another storey taller as vendors competed to become the platform that runs all other platforms.
Here is what you need to know.
Verizon CEO: AI Will Replace “A Large Percentage” of Contact Centre Roles
Execs tend to speak carefully about AI and the workforce. However, at the Bloomberg Tech Conference in San Francisco earlier this month, Verizon’s chief executive, Dan Schulman, told his audience that AI will displace a substantial proportion of the company’s customer service work. He said he could not see how any executive could look their employees in the eye and suggest otherwise.
The remarks arrived with rather inconvenient evidence attached. After three months of running AI agents alongside human customer service representatives, Verizon reports a customer satisfaction rate 1,280 basis points above its previous benchmark. The routine interactions go to AI. Complex ones are handled jointly by the human and the agent. The customer, according to the data, prefers it.
Forrester added its own numbers to the conversation this week, forecasting that 49% of customer service jobs will vanish by 2030. Gartner is more circumspect, expecting at least half of planned AI-driven headcount reductions to stall by 2027 due to data quality and implementation complexity. Buyers would do well to note both.
Read CXM’s full analysis of the Verizon CEO’s AI and contact centre workforce comments.
Oracle’s Q4: A Record Quarter, a $40 Billion Capital Commitment, and 1,000 Agents Shipped
Oracle’s fourth quarter was not short on material. Revenue reached $19.2 billion, up 21% year-on-year, with cloud infrastructure growing 93%. The company delivered more than 1,000 AI agents across its application suite in the past year. These are embedded in core workflows rather than bolted on as sidebars. Over 300 Fusion customers went live during the quarter alone.
The headline that moved the stock in the wrong direction was the capital plan. This was $40 billion in debt and equity in FY2027, including a $20 billion equity offering, to fund data centre construction. Free cash flow was negative $23.7 billion, and capex jumped 162%. Oracle’s remaining performance obligations, its contracted future revenue, reached $638 billion, up 363% year-on-year. This, naturally, substantially reframes the commitment.
The development most relevant to CX buyers is the rollout of outcome-based pricing across the full application portfolio. Fees tied to measurable results rather than seat counts change the internal approval dynamic considerably. The CFO conversation becomes about a demonstrated outcome rather than a capability claim. Thirty-three token bundle customers are still a limited number, but the direction of travel is clear.
Read CXM’s full analysis of Oracle’s FY2026 results and their implications for CX.
Adobe Quadruples CX AI Revenue and Ditches Seat-Based Pricing
Adobe reported that AI-first revenue inside its Customer Experience Orchestration business grew fourfold year-on-year in Q2 fiscal 2026. It also moved its agentic CX product, CX Enterprise Coworker, into general availability this week. The system watches performance signals across channels, measures them against a defined goal, and adjusts workflows autonomously, with teams retaining approval rights before anything goes live.
The pricing model is outcome-based and standalone, with no per-seat component. Adobe joins Oracle and others in a structural shift that is becoming increasingly difficult for seat-model vendors to ignore.
Across the company, AI-first ARR tripled to over $500 million. Experience Platform revenue grew over 30% year-on-year. Additionally, enterprise customers spending over $10 million annually grew more than 20%.
Read CXM’s full analysis of Adobe’s Q2 2026 earnings and agentic CX strategy.
NiCE Rebuilds Its Platform Around Agentic AI Rather Than Adding It On
At NiCE World 2026 in Orlando, NiCE announced that it has restructured its CX platform around agentic AI as a core rather than a feature. NiCE Cognigy provides the reasoning engine. The Agentic Engagement Plane handles orchestration across customer, employee, and third-party AI agents. Guardian AI governs compliance in real time. Agentic Analytics sits above the stack as a discovery layer.
Two further launches accompanied the platform announcement. The Workforce Empowerment Suite unifies management, quality, coaching, and compliance for human and AI agents under one operating model. NiCE Labs is a dedicated research, benchmarking, and prototyping unit whose purpose is to close the gap between what AI can demonstrate in a controlled environment and what it can actually sustain inside an enterprise.
The competitive observation worth making is one NiCE would likely endorse. A great deal of the contact centre market is now competing to be the operating system for AI, rather than a supplier to it. Whether enterprise enthusiasm for agentic architecture is running ahead of the delivery evidence is a separate question. It is also one that Gartner has already asked, forecasting that over 40% of agentic AI projects will be cancelled by the end of 2027.
Read CXM’s full analysis of NiCE World 2026 and the agentic platform rebuild.
Microsoft Adds Conversation Orchestration to Dynamics 365
Microsoft has unveiled Conversation Orchestration, a new capability within the Service Operations Agent in Dynamics 365 Contact Center, now in public preview. The product addresses a structural weakness in contact centre routing: that the initial queue decision largely stands regardless of what changes in the intervening minutes. An agent goes offline, a high-value customer’s wait climbs past ten minutes, and then a transferred call resets its position. The platform simply does nothing, because it has not been told to.
Conversation Orchestration addresses this through natural-language playbooks. Operations admins write instructions directly into the platform in plain language. The system then evaluates these instructions against live CRM data and acts upon them. Dynamic Prioritisation updates customer priority scores continuously. Overflow Based on CSR Availability replaces fixed-timer logic with a live trigger.
The natural-language configuration angle reduces the technical barrier meaningfully for operations teams historically dependent on IT involvement for routing changes. The caveats for buyers are practical. The capability requires a Power Platform pay-as-you-go plan and an Azure subscription. It is not a free Dynamics 365 add-on, and it is currently US-only. The international timeline is unconfirmed.
Read CXM’s full analysis of Dynamics 365 Conversation Orchestration.
Sprinklr Launches AI Brand Visibility Tool — Because ChatGPT May Be Misrepresenting You
Sprinklr has launched LLM Insights, now in limited preview and targeting general availability in Q3 2026. The tool tracks how a brand is represented across AI search platforms, ChatGPT, Gemini, Perplexity, and connects those findings directly to content and engagement workflows so teams can act without leaving the platform.
The product’s key design claim is that it generates prompts from real customer conversations already flowing through Sprinklr. This is instead of synthetic or keyword-derived queries. The argument is a credible one, and the gap it is trying to close is real. Early adopters this spring found that AI platforms were misrepresenting their products as higher-cost alternatives. It surfaced competitor brands more prominently and reinforced inaccurate pricing from third-party domains. The brands had no visibility over any of it.
AI brand visibility is an authentic emerging category, with specialist tools and broader platforms competing for the same enterprise budget. For CX and marketing leaders, the more pressing question is internal. Who actually owns this problem? It sits awkwardly across marketing, digital, and CX functions, and in most organisations has no single owner. LLM Insights does not resolve that question, but it does, at least, make the problem legible.
Read CXM’s full analysis of Sprinklr LLM Insights and the AI brand visibility market.
Decagon’s Autopilot: An AI Agent That Improves Other AI Agents
Arguably, the most conceptually interesting launch of the week came from Decagon, an AI-native customer service platform. Its new product, Duet Autopilot, is an AI agent whose specific function is to find faults in other AI agents. It drafts the fixes, tests them, and sends the result to a human for sign-off.
The self-referential logic is, frankly, either brilliant or unsettling, depending on your tolerance for recursion. Autopilot also runs on itself, so every reviewer rejection feeds back into how it operates next time. Decagon built a proprietary benchmark, DuetBench, to measure agent self-improvement end-to-end. Autopilot passed 93% of diagnostic tasks, above the average human score on the same assessment.
The human-in-the-loop design is sensible and mirrors a broader pattern across the market. Early enterprise users in financial services and retail report that reviewing conversations at scale by hand is no longer viable at the volumes they are running. Whether the review rejection rate in production proves the tool’s quality, or its limitation, is the number worth watching.
Read CXM’s full analysis of Decagon Duet Autopilot and agentic AI self-improvement.
Get in Touch
That’s your customer experience roundup for the week ending 13 June 2026. If you have CX stories to share, connect with me on LinkedIn or drop me a line at [email protected].
