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.
Our main story today is about why more than half of consumers would rather do almost anything else than contact customer service.
Today’s headlines are from:
Consumers Would Rather Avoid Contacting Support
Genesys published the 5th edition of its State of Customer Experience report this month, surveying 5,811 consumers and 1,560 CX and business leaders across more than 20 countries.
The headline finding is blunt:
More than half of consumers say they would rather do almost anything else than contact customer service.
85% say a poor experience has caused them to spend less or stop doing business with a brand.
The gap the report keeps circling back to is memory.
- 95% of consumers expect information to be remembered across channels, yet 48% of organizations still do not automatically pass information between virtual agents and Support Specialists
That disconnect shows up directly in what customers are willing to tolerate - 84% will give a virtual agent 3 attempts or fewer before giving up on it, and 66% say they would switch to a competitor after three or fewer bad experiences with a brand
- For 21% of consumers, a single bad experience is now enough to send them elsewhere, up 24% from last year’s report
Expectations keep climbing too.
- 92% of consumers expect every organization to deliver an experience on par with the best one they have ever had, regardless of industry or company size
- 91% say a company is only as good as its customer service, a 9-point jump from the prior year’s report
Consumers are not rejecting AI outright.
- 46% say they are comfortable with AI making decisions on their behalf if it improves speed and resolution
Where AI still has to prove itself is orchestration: connecting virtual agents, human agents, and the data both rely on, so a customer does not have to repeat themselves when a conversation moves from a bot to a person.
CX leaders seem to agree that is the direction the work is heading.
- 82% expect autonomous AI agents to orchestrate the customer experience within 3 years
- 40 percent of organizations report they are already using agentic AI in some form
Agentic AI adoption alone will not close the experience gap. What closes it is how well an organization connects AI, human interactions, data, and systems across the full customer journey, which is a heavier lift than deploying a chatbot and calling it done.
Intercom Gives Fin Memory Of Past Conversations
News Summary
Fin (Intercom), a customer service platform combining shared inbox, ticketing, and AI powered Agents, announced that Fin now retains memory of past conversations. Returning customers continue where they left off, and their context carries across channels rather than resetting with every new conversation.
Operational Impact
Support Specialists no longer need to re-ask basic details when a customer returns with a related issue, since Fin already has what was covered in earlier conversations. The same memory travels from chat to email, so a customer who started an issue in one channel and follows up in another does not have to start over, which cuts down repeat explanations for both the customer and any Support Specialist who picks up the thread.
Implementation Considerations
Teams should confirm what counts as past for Fin’s memory window and whether older, resolved conversations should stay in scope or age out, particularly for accounts with a high volume of repeat contacts. Support Specialists reviewing Fin transcripts should also watch for cases where carried-over context is outdated, such as an account change or resolved billing issue, since inherited context that is no longer accurate could steer a new conversation in the wrong direction.
Gorgias Opens Gaia Beta To All Accounts, Adds Reviews In Chat
News Summary
Gorgias, a customer service platform built for ecommerce and Shopify brands, announced two updates. Gaia, an AI teammate that works inside Gorgias, is now in open beta for every account, and it can investigate tickets, search the queue in plain language, and tag or assign conversations. Gorgias also began showing shoppers product review ratings and counts directly inside AI Agent chat, pulled automatically from Shopify metafields.
Operational Impact
Support Specialists get a teammate that can handle plain-language queue lookups and routine tagging or assignment work directly inside Gorgias, cutting into the manual sorting that usually eats into ticket-handling time. On the customer-facing side, shoppers who ask about a product during a chat conversation can now see review ratings without leaving the conversation or opening a new tab, which can shorten the path from question to purchase decision.
Implementation Considerations
Gaia is in open beta, so teams should treat its ticket-handling suggestions as a starting point that still needs review rather than a finished workflow, particularly around tagging and assignment accuracy. The review-ratings feature depends on Shopify metafields being populated correctly, so stores with incomplete review data may see gaps or missing ratings in chat until that data is cleaned up.
Help Scout Brings Company-Wide Support Into View
News Summary
Help Scout, a shared inbox platform built for SaaS and growing support teams, announced Customer Portal, a feature that gives multi-contact companies a shared view of their support conversations. Selected contacts can see what colleagues at their company have already raised with support, check for new responses, and reply themselves.
Operational Impact
Support Specialists handling B2B accounts with multiple contacts no longer have to explain the same status update to several people from the same company, since those contacts can now see the conversation history themselves. This should also cut down on duplicate tickets from different people at the same company reporting the same issue without knowing someone already raised it.
Implementation Considerations
Customer Portal is available on Plus and Pro plans, so teams on Help Scout’s other tiers will need to confirm eligibility before rolling it out. Access is granted per contact rather than automatically, so admins will need a plan for who at each account gets visibility, and teams should think through what conversation history is acceptable to expose across a company’s own contacts before turning it on broadly.
Zendesk Folds Intelligent Triage Into More Plans
News Summary
Zendesk, a customer service platform used across support, sales, and IT teams, announced that Copilot Intelligent Triage, previously a paid add-on, is now included in Suite and Support Professional plans and above. The feature automatically classifies incoming tickets by topic, sentiment, language, and key entities.
Operational Impact
Support Specialists on Professional-tier accounts get automatic ticket classification without their organization needing to buy a separate add-on, which means teams that skipped Intelligent Triage due to cost can now put it to work identifying trending issues and routing tickets by topic. Team leads gain access to AI-generated workflow recommendations that were previously gated behind the add-on pricing.
Implementation Considerations
Teams should confirm their exact plan tier, since the change applies to Suite and Support Professional and above rather than every plan. Classification accuracy depends on ticket volume and language mix, so teams new to Intelligent Triage should plan a review period to check that topic and sentiment tags match how the team already categorizes issues before relying on the automated routing. Note: this update comes from Zendesk’s monthly product roundup rather than a dated changelog entry, so its exact release date within the window is unconfirmed.
Sierra Agents Learn To Navigate IVR Systems
News Summary
Sierra, an AI agent platform for enterprise customer service, announced that agents built on its platform can now navigate IVR phone systems without custom setup. Agents handle multi-level menus, keypad tone prompts, and hold music, and recognize when they have reached a live person rather than a recording.
Operational Impact
Support and operations teams that rely on Agents calling other companies, such as checking a prior authorization with an insurer or a repair estimate with a shop, get a meaningful jump in success rate. In Sierra’s own benchmark, enabling the feature raised the pass rate for reaching a person or completing a task from 57 percent to 85 percent, compared to an agent working from the goal alone.
Implementation Considerations
This capability applies to voice-agent workflows that call external phone systems, so it will matter most to teams already running or considering Sierra’s Horizon agents for outbound and inbound calling. Teams should still expect some IVR trees to behave unpredictably, since Sierra notes that broken phone trees and account-specific data requirements can still block a call regardless of navigation skill, and any workflow built on this should include a path for the agent to schedule a human callback when it hits a dead end.
Ada Adds Audit Log And MCP Connection Visibility
News Summary
Ada, an AI customer service platform, announced a set of security and governance controls aimed at teams running Ada through a formal security review. The update adds an audit log with interface-level attribution, bulk end-user data deletion with job tracking, and a page listing every AI-assistant connection to the account.
Operational Impact
Teams that need to answer security review questions about their AI agent now have a direct record to point to instead of piecing one together manually. The audit log tracks configuration changes, API activity, and test runs, and notes whether the change came through the dashboard, the API, or an AI assistant connected over MCP, which matters as more tools get connected through that protocol.
Implementation Considerations
The MCP connections page is worth checking soon after rollout, since it surfaces every AI-assistant connection already authorized on the account, which may include connections a team forgot about or didn’t know existed. Bulk end-user deletion is capped at 1,000 end users per job, so teams with larger deletion requests will need to plan for multiple batches.
ChurnZero’s Retrospective Agent Drafts Churn Analysis
News Summary
ChurnZero, a customer success platform, announced Retrospective, an AI agent that automatically analyzes an account’s full history as soon as it is marked churned. The agent classifies the churn cause against the company’s existing churn-reason set and assigns a confidence score to its analysis for a customer success manager to review.
Operational Impact
Customer success managers get a first-pass churn writeup pulled from notes, meetings, surveys, and other signals instead of starting the documentation from a blank page after a loss. CS leaders should also get a faster read on churn patterns across accounts, since the classification runs automatically rather than depending on each CSM tagging a cause consistently by hand.
Implementation Considerations
Since Retrospective classifies against the company’s existing churn-reason set, the output is only as useful as that taxonomy, so teams with a stale or overly broad reason list should expect to revisit it before the automated classification adds much value. CSMs should treat the draft as a starting point to correct and save rather than a final record, particularly for churns with mixed or ambiguous causes.
Five9 Connects Call Events To Regal’s Voice Agents
News Summary
Five9, a cloud contact center platform, announced an integration with Regal that connects Five9 call events to Regal’s autonomous voice AI agents through the Five9 AI Agent Connect program. A Five9 call event can now trigger an automated text message or an outbound AI-agent call from Regal, with call data syncing in both directions.
Operational Impact
Contact center teams can set up automatic follow-up outreach after a call ends, such as a text or a callback attempt, without a Support Specialist manually queuing the follow-up. Call records, including campaign, duration, and outcome, flow into Regal as calls complete, so the follow-up agent has the context of what happened on the original call.
Implementation Considerations
This is a third-party integration available through Five9’s AI Agent Connect program rather than a native Five9 feature, so teams will need a separate Regal relationship to use it. Teams should also define clear rules for what triggers a follow-up call versus a text, since an AI agent placing an unexpected outbound call shortly after a support interaction could read as intrusive if the triggering logic is too broad.
GitBook Teaches Coding Agents Its Documentation Syntax
News Summary
GitBook, a documentation platform built for docs-as-code teams, announced Agent Skills, a set of skill files that teach AI coding assistants such as Claude Code, Cursor, and Codex how to correctly edit GitBook documentation. Teams install the skills as a package or download them locally, and the assistant gains a syntax reference for GitBook’s custom blocks, configuration files, and frontmatter.
Operational Impact
Support and knowledge teams that maintain their help center or internal documentation through GitBook’s Git Sync workflow can now let a coding agent draft or restructure docs directly in the repo, with the agent producing valid GitBook syntax instead of plain markdown that needs manual reformatting. This fits teams already treating documentation as code, where changes go through commits and pull requests rather than a separate editor.
Implementation Considerations
GitBook’s own documentation recommends reviewing AI-generated content for unclosed custom blocks, invalid YAML in frontmatter, and broken internal links before merging, so this does not remove the need for a review step. Teams need to create an access token for the agent to interact with GitBook directly, and if an assistant caches instructions, the skill file may need a session restart to take effect after an update.
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.
