Our main story today is about preserving human connection in AI-driven customer experience.
Customer Experience News is a weekly newsletter about the most important news and discussions for Customer Experience and Customer Support Leaders.
This is all the weekly news you need in around 5 minutes.
Today’s headlines are from:
A Wall Street Journal study cited by Gartner found that 93% of executives admit their customer experience is broken. They consider it a critical driver of revenue and growth. Most say they still cannot get their heads around it.
That tension opened a session by Don Scheibenreif, Distinguished VP Analyst at Gartner’s CIO and AI Leaders research group, at the Gartner CIO Leadership Forum in Phoenix, Arizona. His argument was not that technology was the problem. It was that organizations are using technology to get between themselves and their customers, and the distance is starting to show.
Scheibenreif introduced a framework he calls “moments of humanity”, any point in an experience where a customer feels seen as a person. He argued that most organizations already know what these moments are. They just have not built systems around them.
He described when these moments tend to occur: when something goes wrong and a customer is met with a system that does not acknowledge their frustration; when they feel vulnerable or overwhelmed; when they need reassurance that someone is handling a problem. A bank client made the point directly. After a fraud attempt on an account, customers do not need a bot walking them through a process. They need a human voice confirming that everything is going to be okay.
To show what these moments look like when they work, Scheibenreif pulled from his own experience. At a Marriott in Vancouver, he used every piece of available automation — app check-in, digital key, room service through his phone. The next morning, a server named Dwayne walked up, introduced himself, and asked for his name. “I felt like I was seen,” Scheibenreif said. He later sent a note to the hotel manager.
Running through O’Hare on a collapsed connection, he received a text mid-sprint: “Take a deep breath. We’re holding your next flight for a few extra minutes.” The message may have been automated. But a human being had designed it for exactly that situation.
He cited Chewy, the pet supply company, as an organization that has embedded empathy into its processes rather than leaving it to individual judgment. When a customer asked to return unopened dog food after her dog died, Chewy told her to donate it to a local shelter and sent flowers the next day.
The response was not a manager override. It was expected behavior.
Against those examples, Scheibenreif placed three companies that moved the other direction.
- Netflix, when it changed its password-sharing rules, automated most of its customer service response. Customers felt unheard. Subscriptions dropped before the company corrected.
- Salesforce, in 2023, had mid-market enterprise customers describing their support experience as being treated like a number. Many moved to HubSpot.
- Deutsche Bahn in Germany introduced kiosk and mobile-first infrastructure that excluded older riders, who found other transportation.
All three recovered. But the question Scheibenreif pressed was why organizations wait for the crisis.
Empathy, as he described it, has three operating layers.
- Emotional empathy is sharing a customer’s feelings.
- Cognitive empathy is understanding their intent and situation — not just saying “I’m sorry” but recognizing what this moment actually means for them.
- Compassionate empathy is taking action on that understanding.
Technology can assist in all three, but cannot carry the weight of any of them independently.
For CIOs, his most pointed argument was that they have more influence over customer experience than they typically claim, given that technology now mediates most customer interactions. He encouraged them to invest in voice-of-the-customer platforms (naming Qualtrics, Medallia, Oracle, and Salesforce) and to treat frontline employee insight as equivalent input.
The two together reveal where moments of humanity belong in a customer journey.
He described California CIO Leanna Cremens Bailey visiting the LA wildfire burn areas in person. Speaking directly with displaced residents, she found they had no idea how to access state services. Her team built the LA Wildfire Portal within 24 hours and iterated on it daily in the weeks that followed.
The research finding he closed with was specific. Gartner studied what actually drives improvement in CX program results. The top factor was not executive commitment statements or accountability structures. It was leadership removing barriers that prevented frontline employees from acting on what they already knew.
A United Airlines crew gave an economy passenger in a middle seat a snack box arranged as a birthday gift, with a note signed by the flight attendants, captain, and first officer. That crew was not acting outside policy. Someone at United had made it easy for them.
Please feel free to listen to the podcast:
Source: Gartner ThinkCast, Gartner CIO Leadership Forum, Spotify
Help Scout Launches SLA Tracking in the Inbox
News Summary Help Scout, a customer support platform for growing businesses, launched Service Level Agreement (SLA) support. The feature lets teams define first reply time and time-to-resolution targets, which Help Scout applies automatically to new conversations based on conditions the team configures. It is available across all user-based plans.
Operational Impact SLAs surface in the agent’s conversation view so agents can identify at-risk tickets without running a separate report or waiting for a manager to flag them. For teams managing growing volume, automatic SLA assignment removes the manual work of sorting and prioritizing inboxes. The visibility creates consistency across the team — every conversation carries a clear timeline rather than relying on each agent’s judgment about what to answer first.
Implementation Considerations Whether the feature produces accurate prioritization depends on how precisely teams define their SLA conditions — vague or overlapping rules will assign tickets to the wrong policy. Teams without baseline data on their actual response times should establish that before setting targets, otherwise they will be measuring against aspirations rather than achievable benchmarks.
ServiceNow Makes AI Native Across All Products
News Summary ServiceNow announced that its full product portfolio now includes AI, data connectivity, workflow execution, security, and governance as standard components — no longer sold as separate purchases. The release also introduced Context Engine, which maps enterprise-wide relationships, asset dependencies, business intelligence, and data lineage that AI agents query in real time to ground their decisions.
Operational Impact Context Engine changes how AI agents handle service requests. Rather than generating responses from a generic model, agents can reference live relationships between systems, assets, and records. The SDK and Build Agent skills launching April 15 allow developers to build on ServiceNow from any IDE — including Claude Code, Cursor, and OpenAI Codex — which removes the barrier for teams that have avoided custom development due to ServiceNow’s historically closed ecosystem.
Implementation Considerations Bundling AI and governance into every package changes the cost calculation for customers who purchased those capabilities separately. Teams will need to audit what they currently pay for and verify whether the included version matches the depth of their existing setup. Context Engine’s value is proportional to the quality of the data it indexes — organizations with fragmented CMDB records or inconsistent data hygiene will see limited return until that foundation is solid.
Kustomer Integrates Ordergroove Subscription Data Into Agent View
News Summary Kustomer, an AI-powered customer service CRM, launched an integration with Ordergroove, a subscription management platform, that pulls live subscription data directly into the Kustomer agent timeline. Agents can view subscription status, upcoming shipment dates, order history, and itemized costs without leaving their workspace.
Operational Impact Subscription support generates consistent ticket volume — unexpected charges, missed shipments, unwanted renewals. Handling those tickets currently requires agents to work across a support tool and a separate subscription management screen. This integration consolidates that context. Agents can pause, cancel, reactivate, skip, or reschedule subscription orders with one-click actions from the Kustomer timeline. Ordergroove’s structured data also feeds into Kustomer IQ workflows and AI agents, so routine requests like “skip my next shipment” can be handled automatically without custom engineering.
Implementation Considerations The integration requires an active Ordergroove account and an API key configured in both systems. Teams using a different subscription management platform get no benefit from this connection. AI automation for subscription actions requires additional setup in Kustomer IQ — it is not active by default after installation. Before enabling full one-click agent controls, teams should confirm which subscription actions currently require supervisor approval and preserve those guardrails.
LiveAgent Adds Call Listening and In-App Alerts
News Summary LiveAgent published its April 2026 product update, releasing a call listening feature in version 5.62 that allows agents to join a colleague’s live call as a silent observer. The update also includes an in-app news panel that consolidates system warnings and alerts into a single location, replacing the previous notification system that occupied dashboard space.
Operational Impact Call listening gives supervisors a live monitoring option without requiring a separate contact center tool or waiting for post-call review. Observers can send notes to the agent on the call during the conversation, which creates a coaching channel in the moment rather than only in debriefs. The in-app alert panel makes it easier to catch integration errors or account issues without navigating away from the agent interface.
Implementation Considerations Call listening requires owner, admin, or custom role permissions. Teams will need to update role assignments before supervisors can use the feature. Organizations subject to call recording consent laws should verify whether silent monitoring requires disclosure to callers under applicable regulations before enabling it. Sending notes to an agent during a live call adds a coaching channel, but also adds a distraction — teams should set clear norms for when and how notes should be sent to avoid pulling agent attention at the wrong moment.
Five9 and Assembled Partner on Blended Workforce Management
News Summary Assembled, an agentic workforce management platform, announced a Select ISV Partnership with Five9 – Five9’s highest partner tier. The agreement also makes Five9 an authorized reseller of Assembled’s AI-powered WFM system, accessible within the Five9 Intelligent CX interface through a pre-built integration.
Operational Impact For contact centers managing a mix of human agents, AI agents, and outsourced staff, forecasting and scheduling across all three has typically required separate systems. This integration puts Assembled’s WFM tooling inside Five9’s interface, so workforce planners can schedule and optimize across a blended workforce without switching platforms. The reseller arrangement simplifies contracting — teams can purchase and bill Assembled through Five9 rather than managing a separate vendor relationship.
Implementation Considerations The integration is pre-built and does not require heavy IT setup, but the value is proportional to how blended the workforce actually is. Teams whose support operations consist entirely of human agents will not see material benefit over a standalone WFM tool. Customers purchasing through Five9 as a reseller should confirm that the arrangement covers the same feature set and support terms as a direct Assembled contract before signing.
Sierra Launches PCI-Compliant Payments for AI Agents
News Summary Sierra, an AI agent platform for customer service, announced the launch of what it describes as the first Level 1 PCI-compliant payment capability for AI agents, verified by the Visa Global Service Provider Registry. The feature allows Sierra-powered voice and chat agents to handle complete payment transactions — including card and ACH processing — within the conversation, without transferring the customer to a live agent or IVR.
Operational Impact Payment transactions are one of the highest-friction escalation points for AI agents. Teams typically interrupt the automated flow to route customers to a secure system or a human agent for anything involving card data. Sierra’s architecture routes cardholder data through separate PCI-certified infrastructure, so the AI agent never processes raw card details. SiriusXM COO Wayne Thorsen cited the capability as enabling payment resolution without live agent transfers. Sierra reports that one financial services client achieved an 85% resolution rate for card activation through the platform.
Implementation Considerations Level 1 PCI compliance covers Sierra’s data handling architecture, but organizations remain responsible for their own compliance posture. Regulated industries — financial services and healthcare in particular — will need legal and compliance review before routing payment flows through an AI agent. Sierra uses predetermined, server-validated sequences during payment mode to address prompt injection risk, but teams should evaluate that architecture against their own security requirements rather than accepting vendor assurances at face value.
Linear Adds Multi-Level Sub-Teams and Project Discussions
News Summary Linear released two organizational features. Multi-level sub-teams allows organizations to build hierarchical team structures up to five levels deep, with each sub-team inheriting workflows and settings from its parent. Projects and initiatives also now support comments in their activity feed, enabling threaded discussions and decision documentation alongside status updates.
Operational Impact For support and operations teams using Linear for work management, sub-teams address a scaling problem that appears as headcount grows or work splits across regions or product areas. Inherited settings maintain workflow consistency without requiring admins to configure each unit independently. Project comments give team leads a place to document decisions and capture meeting outcomes in context — the @Linear mention feature can trigger document updates, description edits, and issue creation directly from a discussion thread.
Implementation Considerations Multi-level sub-teams are available on Linear’s Enterprise plan only. Teams on Starter or Plus plans will need to upgrade to access the feature. Before restructuring around sub-teams, teams should map their intended hierarchy in full, since reorganizing nested teams after the fact can disrupt existing workflow assignments. The @Linear trigger in project comments — which can create or modify issues and documents automatically — requires governance to prevent unintended scope changes in active projects.
Zipchat Code Connects Agents to Live Codebase Answers
News Summary Zipchat launched Zipchat Code, currently in early access, which connects to a company’s GitHub repositories and databases to answer technical support questions in plain English. Responses include the exact file, line, and commit sourcing the answer, along with live account data where relevant, with the stated goal of reducing technical escalations to engineering.
Operational Impact Technical support teams handling developer or engineering product lines frequently reach a limit where agents cannot answer implementation-level questions and must route to engineering. Zipchat Code addresses this by indexing the actual codebase and returning specific, sourced answers rather than documentation summaries. An agent handling a staging environment connectivity error can see the relevant code reference and live account state. Integrations with Jira, Notion, Confluence, Slack, and PostgreSQL extend the answer surface beyond repositories alone.
Implementation Considerations The product is in early access, which means feature completeness and stability should be validated before relying on it for production support volume. SOC 2 Type II certification is listed as in progress, which may be a constraint for organizations with strict vendor security requirements. Answer quality depends directly on the state of the connected codebase — inconsistent documentation, schema divergence between environments, or poor commit hygiene will surface in the tool’s outputs. Teams should also define which question types agents are authorized to answer with this tool versus which should still go to engineering.
Customer.io Launches Workspace-Aware AI Agent and LLM Actions
News Summary Customer.io released an AI Agent built directly into its platform, alongside a feature called LLM Actions. The AI Agent reads the account’s existing segments, campaigns, attributes, and performance history before generating recommendations. LLM Actions allow teams to call any large language model directly within a campaign workflow without webhooks, external services, or engineering support.
Operational Impact The current friction in using AI for campaign work is that most AI tools do not understand the specific segments, suppression lists, and logic already built inside the platform. Customer.io’s Agent starts from the existing workspace state, so a prompt to build a re-engagement journey references actual inactive segments rather than creating new ones. LLM Actions remove the engineering dependency for message-level personalization — teams can generate dynamic content and drive routing decisions from model output, mid-journey. Every paying account receives 100,000 AI credits valid for 90 days at launch.
Implementation Considerations Additional credits cost $10 per 100,000. Teams running LLM Actions in high-volume campaigns should model their credit consumption before launch, since costs scale with send volume. The Agent’s workspace-awareness is only as useful as the workspace it reads — accounts with inconsistent segment naming, duplicate campaigns, or stale attributes will get recommendations that reflect those issues. Teams should treat the first round of Agent-generated suggestions as an audit signal for workspace hygiene before deploying them in production campaigns.
Customer Support Leaders
No one in the company understands what you have to balance:
- Designing where AI can increase productivity
- Coach your team on new tools
- All while maintaining targets
It feels impossible.
You need a strategic CX partner.
You know what needs to change, but you don’t know how to prioritize it.
Gain clarity with 1-on-1 advisory calls.
We pressure-test your priorities, unblock decisions, and build a roadmap you can execute.
No pitch, just a conversation.
