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 10 minutes.
Today’s headlines are from:
Intercom Adds Channel Routing And Notification Controls
Intercom, an AI-first customer service platform, shipped three updates this week: teammates can now handle work by channel — email, Messenger, SMS, or phone — inside a single shared inbox instead of a separate inbox per channel; teams can control which ticket-state changes trigger customer notifications; and workspaces can configure how long Data Connector request and response logs are retained, either 7 or 14 days.
Channel-based routing lets Support Specialists stay in one inbox while still being assigned only the channels they’re trained on, which removes a common tradeoff between specialization and inbox fragmentation. The notification toggle addresses a recurring complaint from customers who get pinged every time a ticket moves between internal states that don’t matter to them, letting teams keep alerts for milestones like resolution while going quiet on the steps in between. The retention setting is aimed at teams whose security or procurement reviews require shorter data windows for connector logs.
Routing by channel only pays off if teams have already mapped which Support Specialists should own which channels; without that mapping, the new option adds a configuration step without changing outcomes. Notification-state changes need a quick audit against existing customer-facing macros and automations that reference the old blanket notification behavior, since every state still starts with notifications on by default. Shortening log retention to 7 days trades away debugging history, so teams that lean on connector logs to troubleshoot integration failures should confirm 7 days is still enough runway before switching from the 14-day default.
Gorgias Rebuilds Sidebar, Adds SSO And Recommendations
Gorgias, an e-commerce helpdesk platform, shipped four updates this week: its AI Agent chat can now suggest similar products with one tap, the ticket sidebar was rebuilt to be more flexible as part of what Gorgias calls Helpdesk 2.0, two new Single-Sign-On options arrived including the ability to fully disable password login, and the AI Agent’s website re-sync frequency changed from once a day to once every 30 days.
The similar-products option keeps shoppers inside the chat conversation instead of sending them back to browse the site, which matters for merchants trying to convert a support interaction into a sale. The rebuilt sidebar gives agents more control over which widgets they see and in what order, a small change that adds up across a full shift of ticket handling. The SSO update and the AI Agent re-sync change both shift control toward admins: SSO for account security, the slower re-sync for how current the Agent’s product knowledge stays.
Merchants who rely on the AI Agent to answer product questions from the live site should confirm what the new 30-day re-sync interval means for their catalog. If prices, inventory, or product details change more often than that, the Agent could be working from stale information until the next sync. Forcing all users through SSO removes the password fallback, so admins should confirm every team member has SSO access configured before disabling it. The sidebar redesign is worth a quick walkthrough with the team so agents know where their frequently used widgets landed.
Customer.io Ships Summer Release With AI Agent
Customer.io, a customer engagement platform, shipped its Summer Release, more than ten new features including Geofences that trigger messages when someone enters or exits a physical location, the ability to bring your own SMS provider (Sinch, Infobip, or Vonage) instead of routing through Twilio, reusable knowledge sources that attach to multiple AI actions at once, and support for multiple mobile apps per workspace.
Geofencing lets support and marketing teams send location-aware messages, like a reminder when someone leaves a store without completing a return, without building custom location logic or bringing in a third-party geofencing tool. Reusable knowledge sources cut down on the copy-paste work of pasting the same reference text into every AI prompt, and make it easier to keep that reference material current in one place. The multi-SMS-provider option matters for teams sending internationally, since providers vary in coverage and pricing by region.
Geofences are limited to Premium and Enterprise plans, so teams on lower tiers won’t have access without upgrading. Location tracking through geofencing also raises the same consent and battery-usage questions as any location feature, worth reviewing against each region’s privacy requirements before turning it on broadly. Bringing your own SMS provider means Customer.io no longer owns the delivery relationship, so teams switching from Twilio should test deliverability and compliance settings with the new provider before moving production traffic over.
Front Redesigns Mobile App With AI Quality Tools
Front, a customer operations platform, shipped a redesigned mobile app this month along with smarter AI quality workflows, stronger collaboration controls, and new partner integrations.
The mobile redesign is aimed at Support Specialists who need to stay responsive outside the office, giving them a faster way to triage and respond from a phone rather than waiting to get back to a desktop. The AI quality workflow additions extend Front’s existing AI review tooling further into day-to-day ticket handling, and the new partner integrations reduce the number of tools teams have to jump between to keep customer work moving.
Front didn’t publish a detailed breakdown of which AI quality workflows or partner integrations shipped, so teams should check Front’s community changelog directly for specifics relevant to their stack before assuming a particular integration is included. Teams that have built workarounds for mobile limitations should revisit those now that the app has been rebuilt, since some may no longer be necessary.
Plain Adds One-Day Ticket History Import
Plain, a customer support platform built for technical teams, launched ticket history import, letting teams bring in their full ticket history — including original dates, message authors, customer emails, agent replies, and attachments — from any legacy provider or homegrown system.
Teams switching support platforms have historically had to choose between losing historical context or spending weeks on manual migration. Plain says the import can complete in as little as a day, which removes one of the bigger blockers to actually switching tools once a team has decided to move.
Import quality depends on how cleanly the source system exports its data, so teams should expect messier history from older or heavily customized legacy tools. Plain describes the import as resumable, which suggests it’s built to handle interruptions, but teams should still plan the migration during a lower-volume period and verify a sample of imported threads before relying on the full history for search or reporting.
GitBook Launches Official MCP Server For Agents
GitBook, a documentation platform, launched an official MCP server with plugins for Claude, ChatGPT, and Cursor, letting connected AI agents create and configure new docs sites, draft content, edit pages, and restructure entire sites.
Documentation and knowledge teams that already use AI tools day to day can now point those tools directly at GitBook instead of copying content back and forth. Every change routes through a change request for review, which keeps a human in the loop before anything publishes.
Setup takes about a minute according to GitBook, but teams should still walk through what permissions the connected AI agent has once it’s connected, particularly for docs sites with sensitive internal content. The change-request review step is only useful if someone is actually reviewing it, so teams should confirm who owns that review before turning this on broadly.
Sierra Launches Context Engine For Long-Running Agents
Sierra, an AI agent platform, launched Context Engine, a system that turns customer relationships into long-running Horizon agents that pursue business outcomes over days, weeks, or months rather than resolving a single conversation and starting over.
Most AI agent deployments treat each conversation as a fresh start, with no memory of what happened last time. Context Engine is built to carry context forward across interactions, which Sierra says makes each subsequent conversation smarter and could reduce how often a returning customer has to re-explain their situation.
Sierra’s post didn’t detail how Context Engine handles incorrect or outdated context accumulated over a long-running relationship, which is worth asking about directly if evaluating this for a high-stakes account. Teams should also clarify what data Context Engine retains between conversations and how that interacts with data retention and privacy policies, particularly for regulated industries.
DeepL Adds Swiss German As Full Language
DeepL, a language AI platform, added Swiss German as a full supported language after a six-month research effort from its linguistics team, letting customers translate directly into and from Swiss German dialects rather than treating them as a variant of standard German.
Support teams serving Swiss German-speaking customers can now get translations tuned to the dialect’s specific vocabulary and structure instead of a generic German translation that may miss regional phrasing. This matters most for teams with a meaningful volume of Swiss customers where dialect accuracy affects how natural a translated response reads.
Teams should test the new language option against real Swiss German customer messages before rolling it out broadly, since dialect coverage can vary in quality even after dedicated research investment. This is a narrow addition — useful for teams with Swiss-market volume, less relevant for teams without it.
Linear Adds Coding Sessions On Mobile
Linear, a product development platform, added coding sessions on mobile, letting teams review code changes, comment on specific lines, and send feedback to AI coding agents directly from the Linear mobile app.
This extends Linear’s existing agent workflows, previously desktop-only, to mobile, which matters for engineering and product teams who want to keep an agent-driven workflow moving when they’re away from their desk, including support engineering teams that triage and fix bugs found in the queue.
This is a workflow extension rather than a new capability, so teams already using Linear’s coding agents will get the most immediate value. Teams that haven’t adopted agent-driven coding sessions yet should evaluate the desktop experience first before treating mobile as the primary way to review agent work.
BoldDesk Adds Real-Time Sentiment Analysis
BoldDesk, a helpdesk platform, launched Sentiment Analysis, which reads ticket conversations in real time and labels each one positive, neutral, or negative so urgent issues surface first.
Sentiment labeling gives teams a way to prioritize a queue by tone rather than relying only on manually set priority levels, which can help catch a frustrated customer whose ticket was filed as low priority by mistake. It also gives managers a way to spot patterns in negative sentiment without reading every ticket.
Automated sentiment analysis can misread sarcasm, brief messages, or industry-specific phrasing, so teams should treat sentiment labels as a signal to investigate rather than a final priority decision. BoldDesk’s own materials point people toward a walkthrough before rolling the feature out, worth doing with real historical tickets to see how accurately it labels them.
My AskAI Adds Human Takeover And Agent Testing
My AskAI, an AI customer service platform, shipped two updates this week: a one-click way for a human to take over any Freshdesk ticket and stop the Agent from replying, plus a new note-reply tag option, and a QA testing tool that lets teams upload a CSV of past tickets, have the Agent answer all of them, and score the replies before anything goes live.
The takeover control addresses a specific moment: something delicate lands in the queue and a Support Specialist needs to step in before the Agent sends another automated reply, and having that be a single click rather than a multi-step override matters when timing is tight. The QA tool moves error-catching earlier in the process, letting teams see how the Agent handles real historical tickets before it’s answering live customers.
Teams should establish who has permission to trigger a takeover and communicate that to the whole team, since inconsistent use could mean some tickets get manual review and others don’t for no clear reason. The QA tool is only as useful as the ticket sample teams upload: a CSV that only covers easy cases will miss the edge cases most likely to trip up the Agent in production.
SparrowDesk Connects Facebook Messenger To Inbox
SparrowDesk, a customer support helpdesk, added a native Facebook Messenger integration that turns Messenger chats into resolved tickets inside the existing inbox.
Teams currently handling Messenger conversations outside their helpdesk — in Meta’s own tools or a separate app — can now bring those conversations into the same inbox as email and other channels, which removes a channel-switching step from an agent’s day.
Teams should confirm how existing routing rules and macros apply to the new Messenger channel before assuming it behaves identically to email. SparrowDesk is a smaller platform, so teams should also check documentation for any Messenger-specific limitations, like message formatting or attachment support, before fully relying on it for that channel.
Freshservice AI Surfaces Similar Tickets For Agents
Freshworks, the company behind Freshdesk and Freshservice, highlighted an AI feature in Freshservice that surfaces similar past tickets the moment a new one is created, giving agents context right away instead of digging through history.
Surfacing related tickets automatically can cut the time an agent spends searching for how a similar issue was resolved before, which matters most for recurring issue types where the answer already exists somewhere in the ticket history.
Teams should check how the similarity matching works: tickets that look similar on the surface can have very different resolutions, and agents should still verify before copying a past resolution wholesale.
Help Scout Launches Native SMS Support
Help Scout, a customer support platform for growing businesses, launched native SMS support, letting teams connect their own Twilio number and manage two-way text conversations alongside email, chat, and social channels in the shared inbox.
Teams that have been managing SMS through a separate tool can now bring those conversations into the same inbox as their other channels, giving Support Specialists one place to work instead of switching between systems depending on how a customer reached out.
SMS requires connecting a Twilio number, so teams without an existing Twilio account will need to set one up first, along with the compliance requirements that come with texting customers in different regions. Message length matters differently in SMS than email — Help Scout shows live character and segment counts while typing, worth building into agent training so replies don’t split into multiple messages unexpectedly.
Now this is:
Strategic Support for CX Leaders.
You’ve got ambitious Support targets and new metrics but you’re not sure what to prioritize first.
The list is long, the queue is getting longer, and you don’t have time to step back and think about CX strategically.
What if you could pressure-test your thinking with someone who’s spent 20 years building Customer Support operations?
No pitch, just a conversation.
