Build better products with our product team.
Right now there is a bad user experience when installing plugins from the marketplace. When we install plugins that have actions/components that were already built from other templates it says to rename them because they need unique names. When we rename the component in the template it shows as if it’s accepted and allows to install the plugin but then it fails with the same error. The only workaround I found so far is to rename the component already installed, delete and reinstall the template. This is not a great experience because I then have to go back to the first plugin installed and update the components and delete the duplicate(renamed) one. It would be great if there was an option to note that it’s installed already so it doesn’t install it or some other version of that to avoid this.This recently happened with the Outlook Plugins I was installing and some of them have many pieces.
We'll add a new UI language setting that controls the UI (navigation, controls, labels) independent of what language a user types, across Assistant (Web & EmployeeWorks) and Search. Citations will show in the original content language, with an option to translate to the query language.
Updating Basic tier to GPT 5.2. Read more about the model upgrade program here.
Brief Me can now see the pictures, not just read the words. The architecture diagram, the revenue chart, the table someone screenshotted — all searchable now, right alongside the text. Works across PDFs, Word, and PowerPoint.
When installing a new version of the same plugin, either from the marketplace or from a lower production environment, you can now selectively keep previously imported components – reducing duplication & simplifying updates.
Implement a feature that allows customization of the default response displayed when a sensitive question is detected. The current message—"I am sorry that I can't help you."—may not be sufficiently informative or helpful for users dealing with serious inquiries or issues.We recommend enabling the ability to replace this default message with a more meaningful, context-appropriate custom response that can better guide or support users.
Provide a configurable option in Knowledge Studio to exclude specific ServiceNow assignment groups from ticket ingestion and analysis used for knowledge recommendations.Certain assignment groups are managed by third-party vendors who maintain their own external knowledge bases. Ticket activity from these groups should not be treated as internal knowledge gaps.Knowledge Studio currently includes all ticket data, resulting in non-actionable and irrelevant recommendations from vendor-managed queues.Recommended solution: Enable assignment group–level filtering (denylist) to exclude selected groups from:Ticket ingestion Knowledge gap analysis Recommendation generation
Request to implement a different color theme for the sandbox environment to ensure clear visual distinction from the production environment and reduce risk of confusion.
Making complex AI reasoning easier to digest and more transparent
InspirationApple Siri AI announcement (June 2026) introduces automatic mode detection for user queries. Simple searches default to Spotlight search results; complex requests route to LLM-powered explanations. This pattern creates a seamless unified experience without user friction or explicit mode selection.Problem StatementUsers interacting with Moveworks currently experience cognitive load when determining whether they need quick search results or deeper AI-assisted reasoning. This split mental model across different bot capabilities or interfaces reduces perceived simplicity and increases support friction for end-users.Proposed SolutionMoveworks implements a unified bot interface with intelligent query routing:Single entry point: Users always interact with one Moveworks bot.Automatic mode detection: The bot analyzes incoming queries and determines the optimal response pathway.Response modes:Enterprise Search mode (simple/direct queries): Returns curated knowledge base results, IT catalog items, HR policy documents, or employee directory records with minimal latency and maximum precision. AI Assistant mode (complex/nuanced queries): Leverages LLM reasoning for contextual analysis, multi-step guidance, or explanatory responses that benefit from natural language generation.No friction: Users do not select a mode or mode type. The bot makes this determination transparently.Business ValueUnified experience: Single bot builds user confidence and reduces learning curve across the organization. Reduced support load: Lightweight queries get fast answers without LLM overhead. Complex queries benefit from AI reasoning without forcing structured search workflows. Improved adoption: Employees default to the bot as a reliable all-in-one agent rather than navigating separate search tools or AI assistants. Operational efficiency: Query routing optimizes cost and latency by matching query complexity to appropriate infrastructure.Technical ConsiderationsClassification logic: Implement lightweight pre-processing to categorize queries (intent detection, keyword heuristics, query length, entity recognition). Avoid excessive latency overhead. Fallback behavior: Define graceful escalation if classification is ambiguous. Analytics: Track mode distribution, user satisfaction by mode, and accuracy of routing decisions to refine classification thresholds over time. Transparency (optional): Consider subtle UI affordance (e.g., "Search results" vs "AI insights" label) to help users understand why they received a particular response type without forcing explicit selection.Competitive PositioningApple Siri AI demonstrates this pattern as table stakes for modern AI assistants. Moveworks adopting this approach positions the platform as intuitive and frictionless, matching user expectations set by consumer AI products while maintaining enterprise-grade knowledge integration.
SummaryMulti-line textbox (Description field) in Web Assistant adaptive cards cannot be resized. Impacts usability vs expected behaviour in modern web forms. Problem StatementUsers cannot expand the Description field. Limits visibility of typed content. Creates friction when entering longer responses. Behaviour differs from other experiences (e.g. Teams variant handled separately but has the same issue). Current BehaviourFixed-size textbox in Web Assistant adaptive cards. No drag-to-resize or auto-expand capability. Expected BehaviourTextbox supports: Manual resize (drag handle) Auto-expand based on content length Scroll fallback only after max height reached ImpactReduced user experience in Moveworks portal flows. Slower form completion for Service Desk requests. Increased risk of truncated or poorly formatted input.Business ContextUsed in bot-fillable forms via Moveworks → ServiceNow integration. Affects high-volume request scenarios (e.g. Ad Hoc Request forms).Proposed SolutionEnhance adaptive card rendering in Web Assistant: Enable dynamic height for multi-line text inputs Add optional config flag for “resizable = true” Maintain parity with Teams or other supported channels where applicable Alternatives ConsideredNone viable currently (confirmed by Engineering as enhancement only). Success CriteriaUsers can comfortably enter long-form text without UI constraints. Improved completion speed and quality of inputs. Reduced agent follow-up due to incomplete descriptions.CurrentFuture State - What it could look like
Implement activity logging to capture and track user access to the Interactions table within the Analytics module. This enhancement would enable visibility into who viewed the data, when it was accessed, and potentially from where, supporting auditability, security monitoring, and compliance needs. The goal is to ensure that all access events are recorded and available for review, helping teams maintain transparency and control over sensitive interaction data usage.
Updating our Reasoning Engine to GPT 5.4.
SummaryMoveworks in-bot fillable forms currently default to US date format (MM-DD-YYYY) and do not respect user or tenant locale settings such as en-AU (DD-MM-YYYY).ProblemDates appear in US format regardless of regional settings Creates ambiguity (e.g. 04/08/2026) for AU/NZ users Issue is consistent across MS Teams and Web AI Assistant ServiceNow correctly honours locale, highlighting inconsistencyExample below, date is in US format. Investigation OutcomeConfirmed by Moveworks Support and Engineering No configuration option to control date format Behaviour likely tied to rendering layer, not user settings No workaround available todayImpactConfusing user experience Risk of incorrect date submission Additional validation effort for Service Desk Reduced trust in bot-driven request accuracyRequestIntroduce support for locale-aware date formatting in in-bot forms, including:Automatic detection of user locale (preferred) Tenant-level locale override Optional field-level format control (DD-MM-YYYY, MM-DD-YYYY, etc.) Consistent behaviour across Teams, Web, and other platformsPriorityMedium–HighAdditional ContextThis is a basic expectation for global organisations. Lack of locale support is a gap for AU deployments and limits adoption of in-bot forms.Expected OutcomeDates displayed and submitted in Moveworks forms align with user regional settings, removing ambiguity and improving accuracy.
UCA -Preflight Checks ensures uninterrupted, successful UCA plugin runs.We do this in 2 ways:We ask for all of the UCA connectors required from the user before the plugin runs - no surprises. We continue the plugin execution with all of the prior context in tact so that users don’t have to manually “retry” - it just works.
Hello,In Plugin Insights, the section “What content items are being cited by the AI Assistant?” provides a very valuable overview of top-cited content, including counts of helpful and unhelpful feedback as well as related tickets. However, the current experience makes it difficult to take the next analytical step directly.Current LimitationAt present, clicking on the unhelpful/helpful feedback count does not provide access to the underlying conversations. To investigate these cases, users must:Access a different module (Moveworks Setup), which requires additional permissions not available in Analytics Manually search for KB article ID Navigate to the feedback table and filter out relevant feedback based on KB article IDThe same goes for tickets data – you can’t access underlying conversations leading to ticket creation from hereThe only thing you can drill down from this view is total interactions table which does not contain any feedback related data in the table/raw dataCurrent set up limits the ability to quickly derive insights.Proposed ImprovementEnable direct drill-down functionality from Plugin Insights Allow users to click on the helpful or unhelpful feedback count for a given content item Display a filtered list of all content-related conversations with users’ feedback entries
As per: https://help.moveworks.com/docs/multilingual-support-assistant and https://help.moveworks.com/docs/multilingual-support-copilot multilingual support is capable of translating everything EXCEPT an Agent’s comments back to a user’s preferred language on a ticket. A one-way translator is not sufficient to maintain a conversation. This is not a fully completed product as is and should not be advertised as such.PLEASE enable two-way translation for ticket comments.
I have created a couple of highly customized connectors in Moveworks and would like to put them for sell. Is there a Marketplace where I can share it for users. Something equivalent to ServiceNow Marketplace where I have my customized scoped hosted and make revenue in application keys in accordance to its usage.
TL;DRAdd connector visibility in UI → eliminate guesswork across environments.ProblemMultiple ServiceNow environments (DEV → UAT) connected in Moveworks. Forms with the same name appear multiple times with no indication of source environment. When a form fails to fill, it’s unclear which environment it belongs to. Current workaround requires checking ingestion timestamps/logs → inefficient.RequestSurface Connector / Environment on: Imported Forms table (new column) Form detail page (Overview section) Expected OutcomeQuickly identify whether a form is DEV or UAT. Remove need to cross-check ingestion logs. Faster troubleshooting and safer configuration.
Already have an account? Login
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.