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.
This week’s headlines span AI Agent behavior, knowledge management, scheduling, and a handful of new partnerships.
Plain Teaches Ari When to Step Back
Plain, a B2B customer support platform, shipped six improvements to Ari, its AI Agent, this week. The changes focus on knowing when not to respond: handing off to a Support Specialist when a customer tags a teammate in Slack, recognizing autoresponder loops, and catching messages that don’t need a support reply in the first place.
Operational Impact: Each change targets one specific moment where Ari used to keep talking when it shouldn’t have. A customer who tags a colleague in a shared Slack thread now gets a handoff with a note instead of Ari answering over the person they actually asked for. Threads a Support Specialist starts are left alone so Ari doesn’t interrupt a conversation it wasn’t part of. Bot-to-bot loops, well-wishes, and FYIs get redirected instead of producing a reply that reads as tone-deaf, and Ari no longer promises an action like a refund or account change without a real handoff completing first.
Implementation Considerations: These are refinements to Ari’s judgment, so teams already running it should see the change without reconfiguring anything. The open question is how often these situations were happening before the update, teams that already track Ari’s handoff and escalation rates will have a clearer read on the impact than teams that don’t track those numbers at all. The promise-handling fix is worth testing directly: have a Support Specialist run a scenario where Ari can’t complete an action and confirm it hands off instead of falling back into an old pattern.
Notion Adds Calendar Connections for AI Agents
Notion, the AI workspace platform, introduced a calendar connection this week that links a Google, iCloud, or Outlook calendar directly to a workspace. Once connected, Agents can schedule events and write meeting notes automatically, and teams can manage Notion databases from inside the calendar itself.
Operational Impact: Teams that already track tickets, projects, or on-call schedules in Notion lose a manual step between the calendar and the database. An event created from a Notion database now shows up on the calendar without copying it over, and meeting notes land in Notion without someone opening a doc afterward to write them up. Teams running stand-ups, retros, or logged customer calls will feel this most directly.
Implementation Considerations: The feature depends on connecting a real calendar account, so teams split across multiple providers or under strict IT policy on third-party calendar access may hit friction before they see the benefit. Automatic meeting notes are only as useful as how cleanly they get filed. Without a defined home for them in the workspace, they risk piling up the same way an inbox does. Teams should decide where these notes belong before turning the feature on broadly.
Linear Turns Personal AI Skills into Team Tools
Linear, a project and issue tracking platform used by many support and engineering teams, added a way to turn a personal AI skill into a shared team workflow this week. A team member can ask Linear to share a skill they built, and it becomes available to everyone on the team.
Operational Impact: Useful AI shortcuts often stay locked to whoever figured them out first. This gives a team a way to standardize on the skills that already work, instead of every Support Specialist or engineer rebuilding the same workflow alone. A skill built to summarize an escalation thread or draft a release note can now scale across a team without someone copying instructions by hand.
Implementation Considerations: Sharing a skill assumes the original version was built well. A skill that worked for one person’s specific use case may not generalize once a whole team relies on it, and someone needs to own maintaining shared skills as workflows change. Teams adopting this should also decide who has permission to share a skill broadly versus keep it personal, since not every shortcut is ready for team-wide use the first time it works.
Zipchat Partners with Dyrect on Warranty Support
Zipchat, an AI customer service platform for ecommerce, announced a partnership with Dyrect, an AI warranty software company, this week. The integration connects Zipchat’s AI conversations with Dyrect’s tools for product registration, warranty claims, repairs, and replacements.
Operational Impact: For ecommerce brands selling physical products, the relationship with a customer doesn’t end at checkout, and this partnership targets the warranty and claims work that follows. A customer asking Zipchat’s Agent about a broken product can now get routed into Dyrect’s claims and replacement flow instead of being handed to a separate system or a Support Specialist who has to look up warranty status by hand.
Implementation Considerations: This is a partnership announcement between two platforms, not a description of a finished joint workflow at any specific brand yet. The value depends on how deeply a merchant has Dyrect’s warranty data connected and how much of that information Zipchat’s Agent can actually surface mid-conversation. Brands evaluating this should ask what specific warranty actions the Agent can complete versus what still requires a handoff, before assuming the after-sales experience is fully automated.
HubSpot Launches Revenue Hub to Unify Billing
HubSpot launched Revenue Hub this week, a new product that brings quotes, contracts, billing, and payments into one place built directly into the platform. The goal is to keep deal data synced across the revenue process instead of split across separate tools.
Operational Impact: HubSpot’s pitch is that quotes, contracts, and billing living in separate tools is what stalls a deal close, and Revenue Hub keeps all of it inside HubSpot to close that gap. For support and renewals teams that get pulled into billing disputes or contract questions, having that data in the same system as the customer record could cut down on the back-and-forth needed to answer a billing question.
Implementation Considerations: This is a sales and revenue operations product rather than a support tool, so its relevance to a support team depends on how closely support and revenue operations already work together. Teams considering it should weigh what it actually replaces, since billing and contract tools already in place may have functionality or integrations Revenue Hub doesn’t yet match in its first release.
Miro Integrates with Microsoft Copilot Cowork
Miro, the visual collaboration platform, announced an integration with Microsoft Copilot Cowork this week. Conversations in Copilot Cowork can now be turned into a fully built Miro board, ready for a retrospective or team sync.
Operational Impact: Teams that already run retros, planning sessions, or working sessions in Miro get a way to skip the blank-board setup step. Copilot Cowork can also convert team input into structured tables that organize ideas and surface priorities, useful for a team translating a loose discussion into something actionable without someone manually transcribing notes onto a board.
Implementation Considerations: This depends on a team already using both Microsoft Copilot Cowork and Miro, so it adds the most value to organizations with both tools in place already. As with most AI-generated boards, the output is a starting point. A board built from a conversation will likely need a human pass to clean up structure or correct what the model misread before a team relies on it in a live session.
Zendesk Strengthens AWS Partnership with New Competency
Zendesk strengthened its partnership with AWS this week, announcing it is one of the first 13 global launch partners to achieve the new Amazon Connect Customer Competency and being featured in an AWS Partner Network blog post on building what AWS calls Agentic SaaS. Zendesk is also listed in the AWS Marketplace.
Operational Impact: For contact centers already running Amazon Connect, the competency signals that Zendesk’s integration has been vetted against AWS’s own standards for that platform, which can simplify a procurement conversation for teams that need to justify a vendor choice internally. Marketplace availability also means teams already buying through AWS can purchase Zendesk through an existing AWS billing relationship instead of setting up a separate vendor contract.
Implementation Considerations: This is a partner program and marketplace announcement, not a new Zendesk feature, so it changes how a team can buy and validate Zendesk rather than what Zendesk does day to day. Teams not running Amazon Connect or not procuring software through AWS Marketplace will see little practical difference from this update.
Siena AI Launches a Knowledge Layer for Brands
Siena AI, a CX Agent platform, launched Siena Docs this week, a knowledge layer built specifically for brands. The product combines a team’s internal wiki, customer-facing help center, and the source material an Agent runs on into one synced layer, with built-in verification and analytics on what’s actually being used.
Operational Impact: Documentation scattered across a help center, macros, a website FAQ, and a stray PDF is a familiar problem, and it gets worse once an Agent is grounded on all of it at once. Siena Docs lets any page carry an owner and a status, so a team and an Agent both know what’s confirmed versus what isn’t, and any page can convert into a Skill the Agent runs on directly. The analytics layer shows which pages a team, an Agent, and customers each actually use, which gives a knowledge manager a real signal for what to update or archive instead of guessing.
Implementation Considerations: A knowledge layer is only as good as the migration into it. Teams with documentation spread across five or more tools should expect a real consolidation effort before Siena Docs becomes the single source it’s designed to be, not an instant fix once it’s turned on. The verification and ownership fields also require someone to actually assign and maintain them, an unowned page with an unclear status defeats the purpose of adding that field in the first place.
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.
