Introducing 2026.1!
Deskpro 2026.1 focuses on making AI more practical, more controllable, and better aligned with how your teams actually work, while continuing to improve reliability across the platform.
So, what’s new in 2026.1? Here are the highlights:
tl;dr – What’s New? 🚀
Here’s a quick summary of the major changes. Click and jump into a specific section for more detail:
Snippets as an AI Content Source - AI suggestions, powered by your Snippets.
Suggest a Reply improvements - Smarter Suggest Reply, with admin-managed contexts.
Berget AI - Connect Deskpro Private AI to Berget AI and run models like Mistral on European-hosted infrastructure.
Deskpro AI:
Snippets as a content source
Deskpro’s suggest reply feature can now use Snippets as a source of data. This allows AI suggestions to be grounded in the trusted, real-world responses your agents already use, while fully respecting existing snippet visibility and permission rules.
Available to Professional and Enterprise customers.
Suggest Reply improvements
Agents can now use simple conversational instructions in the Suggest Reply chat to confirm a response, move back to a previous suggestion, or ask the AI to summarise their instructions and context. This makes refining replies faster and more intuitive, without relying on rigid commands or extra clicks.
Alongside this, admins can now review, manage, and create reply contexts in one place. This gives admins clearer visibility into how Suggest Reply is being used and helps maintain consistency across shared and global contexts.
Berget AI
Deskpro Private now supports Berget AI as an AI provider. Customers can connect to Berget AI and select from available models.
Berget AI provides a European-hosted, OpenAI-compatible platform with access to open-source models like Mistral, making it easier to run Deskpro AI on privacy-focused, sovereign infrastructure, without changing existing AI workflows.
Polish & Fixes ⚒️
AI
Fixed an issue where the ticket summary field could become read-only after generating, saving, or reopening a summary.
Improved the visual consistency of dropdown menus when adding AI connections.
Fixed an issue where adding or updating a data source triggered unnecessary re-indexing of all existing content.
Fixed an issue preventing AI draft replies from being generated when using Claude via AWS Bedrock.
Improved how AI features determine the most recent user message for context.
Improved the AI Web Sources feature to cache robots.txt checks, reducing unnecessary requests and improving performance.
Improved the reliability of AI web scraping to ensure indexed content matches the source website.
Improved the reliability of AI Web Source indexing to ensure large crawls are completed as expected.
Improved AI Web Sources to better detect and log timeout and crawl failure errors.
Improved AI web scraping security by validating all DNS records to prevent access to private or internal IP addresses.
Tickets
Fixed an issue where the ticket summary field could become read-only after generating, saving, or reopening a summary.
Fixed the navigation to avoid unnecessary re-rendering of ticket queues.
Fixed an issue where creating tickets from templates could fail silently, and template fields were not saved or displayed correctly.
Fixed an issue where Ticket history could fail to load on tickets created via Messenger.
Fixed an issue where selecting Download from the attachment hover menu opened the file instead of downloading it.
Fixed an issue where closing the last ticket tab in dual view caused the ticket list to scroll back to the top.
Fixed an issue where tickets without replies were sorted incorrectly when queues were ordered by Date Last Reply.
Fixed an issue where opening the queue preview from the header could fail to display ticket details.
Fixed an issue where tickets could not be created from templates when CC email addresses were included.
Fixed an issue where voice message playback could reset unexpectedly while listening to recordings.
Fixed an issue where long URLs could overflow message bubbles to agents viewing a ticket.
Fixed an issue where line breaks from multi-line custom fields were lost when inserted into replies via snippets, ensuring formatting is preserved in sent messages.
Fixed an issue where failed approval emails could create invalid ticket history entries, causing the ticket history to break.
Fixed an issue where inline images in agent notes were sent as links in email notifications instead of being embedded as images.
Improved SLA queue accuracy so tickets are correctly removed from warning or failed queues once SLAs are completed.
Messenger & Chatflows
Fixed issue where duplicated chatflow nodes would disconnect when linking to other nodes.
Resolved issue where the fallback route configured in the “Pass to Agent” node of a chatflow caused the removal of the chatbot history upon ticket creation.
Resolved input label inconsistencies when mapping chat flow inputs to custom fields.
Fixed an issue where the messenger conversations could appear across multiple brands, ensuring conversations stay within their correct brand.
Fixed an issue where valid links containing hyphens couldn’t be saved in Messenger link cards.
Fixed an issue where the Messenger widget appeared for users who didn’t have permission to use Messenger when it was set to message-only mode.
Fixed an issue where the “Submit a Ticket” card in Messenger would intermittently fail for some users.
Fixed an issue where duplicate agent presence updates were shown in the Messenger widget.
Help Center
Fixed a permission issue that allowed agents without Help Centre access to post comments on knowledgebase articles via the API.
Fixed the issue where the Reject Approval button was no longer appearing in the Help Center.
Fixed an issue where social share icons (such as LinkedIn and Twitter) were not displaying correctly in the Help Centre.
Fixed the issue of the banner image being set incorrectly when a new brand is created.
Fixed an issue where the field picker on the ticket list page could fail to open.
Fixed the issue of the auto-publish date for a Knowledge base article being uneditable once set.
Fixed an issue where existing users submitting tickets via an embedded form could encounter an error when trying to view the ticket.
Improved the reliability of adding and removing CCs on tickets via the Help Centre.
API
Fixed a bug which caused issues when creating Community Topics via API with some URL slugs.
Fixed searching for organizations via the API will now also return the phone numbers of the matching organizations in their list of fields in the GET request.
Fixed the issue preventing admins from editing the API limit values.
Other
Removed legacy "Chat permissions" option in CRM UserGroups if the instance has never used legacy chat
Fixed an issue where agent preference settings could occasionally reset without agents taking action.
Fixed an issue where Legacy Chat email transcripts were sent even though the user had not requested them.
Fixed an issue that could expose sensitive credentials in authentication error logs during Private instance creation.
Fixed an issue where temporary attachments were not being automatically cleaned up, which could lead to unnecessary storage usage.
Updated the in-product help link for report widget overrides to ensure it directs to the correct guidance and examples.
Implemented security and dependency updates to strengthen platform protection.
Improved task permission handling so only permitted agents can comment on private tasks.
Improved MFA security by preventing unauthenticated access to MFA setup endpoints, closing a loophole that could allow MFA bypass during agent authentication.
Fixed the Jira app so that agents can now link tickets to existing issues in Jira.
Fixed SCIM endpoint to return proper 409 Conflict responses with SCIM-compliant error messages when duplicate user creation is attempted
Fixed 2FA from failing due to a missing email address in the configuration. An email must now exist to enable 2FA with email.
Stopped the logging of sensitive information in the request body of failed SCIM requests.
Fixed the CRM Sync so that webhook payload tokens (such as user timestamps) are correctly populated with user data.
Resolved issue with miscalculating long-range SLAs as a result of them pausing inadvertently.
Fixed the issue of not being able to include Hidden custom fields in Escalations and Triggers.
Fixed issue where using a multi-select field in Macros would break when converted to a single-select field
Fixed department dropdowns across admin areas to show full hierarchical labels (e.g. “English > Technical”) instead of just the child name. This makes it easier to distinguish between departments with the same child name when configuring triggers, SLAs, and other rules.
Fixed the Agent Teams page so it now loads with teams sorted alphabetically (A to Z) by default, making it easier for admins to find teams without manually adjusting the sort order.
Fixed an issue where radio button–style select fields could not be used as criteria when creating or editing ticket queues.
Fixed an issue where holiday dates in triggers could display incorrectly due to time zone handling.
Fixed an issue where triggers using Billing charge-based conditions did not run correctly when tickets were updated.
Improved responsiveness when typing in the Voice Queue and Auto Attendant name fields.
Fixed an issue where adding new options to select fields could unintentionally affect existing tickets; admins are now guided to safely delete or replace impacted options.
Fixed an issue where creating a new agent could fail even when seats were available.
Fixed a SCIM sync issue where users with the same email address but different casing could fail to provision due to a duplication error.
Fixed an issue where CRM Sync webhooks only fired for the first user created, instead of triggering for every new user.
Fixed an issue where CRM Sync updates to Salesforce could be skipped because required Salesforce reference fields were not being populated correctly.
Improved the wording for the Maximum collective file size limit in Email settings to make attachment limits clearer.
Added the ability for customers to opt out of product analytics collection.