Product Forums

We had our highest attendance yet, 40 people on the call and the questions did not let up! Ryan was meeting with a customer and Steve was still waking up on the West Coast, so I ran this one solo — and honestly, it turned into one of the most teaching-heavy sessions we've done. I spent the first chunk breaking down how LLM reasoners works under the hood, then we got into live debugging, scoping gotchas, and a sneak peek at the observation interpreter.If you've ever wondered _why_ your plugin works sometimes and falls apart other times, this session was for you. Let's get into it.How the Reasoner Works: Context, Attention & Why Less Is More This was the big one. I walked through how the LLM actually processes everything in Agent Studio — your plugin config, slot descriptions, conversation messages, action outputs — all of it gets packed into a context window as tokens. The reasoner uses attention to figure out what matters, but here's the thing: attention drops off hard in the middl

Hi, We are establishing LiveAgent handoff with message brokering (Servicenow is our ITSM).Is there an option to update the message that bot sends to user?For example, when there are no active agents, we see message as “I attempted to start a live agent chat for you, but there are currently no agents available to handle your request, so you were not added to the queue.”.Do we have ability to update this message to give more instructions like file a ticket etc. ? Also, we won't have any agents at all on Sundays, is it possible to define that if someone tries to reach out on US Sunday or after hours, add a message that they are reaching after hours?

Hi community,Would it be possible to reconsider deprecating the Events section in the Moveworks Classic Agent Studio? At the moment, it’s no longer possible to create new projects using this functionality.In my view, the Events API and listeners serve different purposes and are not interchangeable.The Events API is well suited for simple automation use cases where there are no dynamic variables and minimal complexity - such as sending predefined, global messages. It’s lightweight, easy to configure, and very effective for these scenarios.Listeners, on the other hand, introduce additional features which is perfect if we want to show buttons and do event processing, but also complexity and they have timeout limitations associated with compound actions.For these reasons, retaining the Events functionality would still provide value for straightforward use cases.Thank you for considering this.

Hi, Does Moveworks has the capability to read and process data directly from an attachment?For example: “Please validate whether the attached document is compliant with the organization’s data security policy.”In this scenario, the user would attach the document for assessment. Regards,Rajeev

Hi everyone,I am building a Plugin in Agent Studio that retrieves a user's timecard details and presents them with an "Approve" button using a Compound Action. I am running into an issue where the Reasoning Engine hallucinates a successful approval status immediately upon displaying the card, rather than waiting for the user to click the button.The GoalPlugin triggers (e.g., "I need to approve my timecard").Compound Action fetches data, formats it using LLM, and sends a notify message with an "Approve Timecard" button.The Bot should say: "Here are your details, please review and approve."The IssueWhen the plugin runs, the bot outputs the following message at the top of the response sometimes at the end, before the timecard details:"Your timecard has been retrieved and successfully approved."It seems the AI believes that simply running the plugin successfully means the business process (approval) is complete.My Setup:I am using a Compound Action called via an Action Activity in a Conver

We’re introducing Ticket Viewer, the first troubleshooting tool available in Moveworks Setup. Ticket Viewer helps admins and implementation partners debug ticketing workflows faster by letting you look up a ticket and review the data Moveworks is using for ticket processing.What you can do with Ticket ViewerUse Ticket Viewer to troubleshoot common scenarios like: Confirm the latest ticket status to ensure state mappings are working as expected Verify the “requested for” user to ensure notifications go to the right person Review work notes or comment logs to determine whether a customer-facing comment was added by an agent through activity logs Identify the internal destination mapped to a specific ticket Review user details (such as user tags) to determine whether any DSL rules affect the ticket workflow   🔐 Access and RBAC (New role introduced)Because Ticket Viewer may display sensitive ticket and user information, access is governed by role-based access control (RBAC).To sup

Hi everyone,We’re rolling out a small upgrade that improves how approval conversations are reflected in reporting. What’s improvedYou may now see multiple users associated with the same conversation when approvals are involved. This means that for same conversation ID,  you might see different user IDs.  This gives a more complete and accurate view of who actually participated. Why this is betterIn approvals, there are usually two key participants: the person who requests the approval (requestor) the person who approves it (approver) Since both people take actions in the same conversation, showing both users’ interactions under that shared conversation provides a clearer picture of what happened. If you have any questions, feel free to reply here. Thanks! 

We’re seeing unexpected color changes in Microsoft Teams that are causing usability problems. Link colors are not rendering as selected. Links show blue/black even when green is configured The Teams UI has shifted to a yellow palette, which has poor contrast and looks unpolished in Teams There was no prior communication about this change. This will negatively impact user interactions and trust if not fixed quickly.Can this be reviewed and reverted or corrected ASAP? 

Hey builders!Steve, Ryan, and I were in the kitchen today – we launched Optional Slots live on the call, went deep on passing data between actions, tackled output formatting best practices, and walked through user consent auth with Microsoft Graph.Here's the recap of what we crushed: Optional Slots – Launched Live on This Call 🔥 → Three slot inference policies now available: Always Ask, Infer if Available, and Always Infer. Always Infer means the assistant never prompts for that slot – it falls back to a DSL expression or null. → Two big patterns: smart defaults (fall back to a recently fetched value) and null values (let unused fields pass as null for generic update plugins). → Docs: Optional Slots Passing Data Between Actions – Available Data Bank Panel Demo → The Available Data button in the action editor shows every field available from previous actions and slots. Click to copy, drag and drop. → Set an output key on your action, then reference it downstream with data.output_key.

Hi team, I am building a workflow using the notify action with a webhook listener. I want to send a message to the end user that is preceded by a greeting. When using the classic events, there was an option to use a greeting (Hi, {{name}}). Is this something that we can turn on with the new webhook listeners? If not, what would be the best way to implement this? 

Sharing Solution to HTML-encoded string passed from Webhook call to present URLs to users through Teams Notification (includes mailto: clickable link and URL) Notify action to handle link:steps: - action: action_name: mw.get_user_by_email output_key: mw_get_user_by_email_result input_args: user_email: data.email - notify: output_key: output_notification recipient_id: data.mw_get_user_by_email_result.user.id message: RENDER(): template: "Hello {{name}}{{& message}}" args: name: data.mw_get_user_by_email_result.user.first_name message: data.message.$REPLACE("&", "&").$REPLACE(""", "\"").$REPLACE("<", "<").$REPLACE(">", ">")Markdown Syntax Reference Documentation  Sample JSON Payload to Webhook:{ "event": "employee.created", "data": { "employee_id": "12345", "name": "John Doe", "email": "william@test.com", "department": "🚀 Engineering", "message": "

Hi Moveworks Community,We have 44 external knowledge entities enabled (Zoom, Adobe, Slack, etc.) and I'm trying to understand how often our bot is citing external vendor documentation vs. our internal Corning content.What I've tried:Looking at the "Content items cited by AI Assistant" widget in Knowledge InsightsThe challenge: The content item names in the dashboard don't clearly indicate if they're from internal SharePoint sources or external vendor sites. URLs aren't easily visible in the widget view.What I need:Is there a filter or report that shows "Internal vs. External" content breakdown? Or a way to easily identify which content items are external sources vs. internal? How can I calculate the percentage: X% internal citations, Y% external citations?Is there a simpler way to track this in the analytics platform?Thanks!

I am leveraging optional slots in a conversation process I have been building, but have 2 issues:The conversation failed on setting a slot called conversation_context, that is set to Always Infer, with this error “hit an internal issue trying to build the conversation context and that specific run (execution <execution id>) has been stopped and cannot continue”. This has only happened once so far. I have a user object slot that is getting populated via a notify statement. I have tried all 3 infer options and none of them have been successful in setting the user properly. I can no longer get this functionality to work that was working perfectly fine earlier today.This last issue is big in that it seems to be is breaking notify statements sending a user object to a conversational process.

I have a slot called description where I am asking the user to describe what they are looking for. In general this just ends up being what initiated the plugin. I want to enforce that they provide enough information for it to be handled properly. I see 2 ways to do this currently:Put in a slot validation based upon the overall length, number of spaces, etc Have a second string slot to ask follow up questions#1 does not reliably work as moveworks just adds in additional words to make it long enough 90% of the timeI would prefer not to do #2, as it is possible the user could be providing enough information so additional prompting may not be necessary.Is my best bet just having a secondary slot or does anyone have good ideas to handle this reliably?

Introducing Optional Slots in Agent StudioWe're excited to share a new capability in Agent Studio that gives you much more control over how your conversational plugins collect information: Optional Slots.The problemUntil now, every slot declared in an Activity or Decision Policy required a value before the plugin could move forward. That meant users were prompted for every single field if no value was explicitly provided when NULL, a previous step, or a sensible default would do. For plugins like quick updates, lookups, or one-shot transactions, this added unnecessary friction to the conversation.What's newOptional Slots introduce two key concepts: More robust Slot Inference Policies and Fallback Values.Slot Inference Policies let you control how each slot's value gets collected:Always Ask: The assistant always prompts the user, even if a value could be inferred from context. Infer if Available: The assistant tries to infer the value from conversation context first. If it can't, it pro

When a ticket is intercepted by the bot, the user receives a notification saying “Please check this information, it might help with your ticket.”The response includes a link, but that link is not served or accessible within the ticket context.In the classic version ticket interception, the same link appeared and worked directly from the ticket, which helped agents quickly see what information was shared with the user.It would be great if this functionality could be restored in the Copilot version, so that the link is available within the ticket context as well.

Hi Moveworks Admins,We’re rolling out an enhanced Role-Based Access Control (RBAC) model for E-Comms for our US, EU, AU and CA customers. This update introduces a more granular role structure that clearly separates access, publishing and user management, making E-Comms permissions simpler to manage and more scalable as usage grows. Support for GovCloud is coming soon.Previous E-Comms Permission ModelPreviously, E-Comms used two roles, along with a separate publishing control:Employee Communications AdminFull access to E-Comms Can view, edit and delete campaigns Can manage user access (add/remove users, update roles)Employee Communications UserCan access E-Comms Can view and edit campaignsHow Publishing Worked PreviouslyPublishing was controlled by an additional setting called Access control: Publish Campaign (found under Setup → Chat Platforms → Notification Settings → Additional Controls).A user could publish a campaign only when both of the following conditions are met:The user had e

Hi Moveworks Community — we’re looking for help troubleshooting routing/trigger conflicts and candidate list selection behavior for an Agent Studio software lookup experience. I’m including sanitized configuration details below (no internal URLs, no people info).1) What we built (use case + goal)We built an Agent Studio plugin to let users search an approved software SharePoint list by software title, including spacing/variants and misspellings (e.g., “Power BI” vs “PowerBI”). Output should show Title + Approval Status + (Restrictions only when applicable) + Corning Standard + Order/Access URL.2) What Moveworks engineering has already done (important context)We worked directly with a Moveworks engineer and have already implemented multiple changes/guardrails:Added steerability / deprioritization for forms to reduce surfacing of software request forms in these flows (forms may still be evaluated, but the goal is to prevent forms from becoming the final answer). Implemented a Dynamic Slo

If you're exploring Agent Studio, you'll come across plugins. These aren't your typical software components. Think of them as skills you can teach your AI agents to do more, move faster, and scale across your enterprise.Our Staff Developer Advocate ​@Kevin Mok breaks it down in under a minute 👇Video linkWant to go deeper? Join Kevin for developer office hours this Thursday at 1 PM PST: Register here.

Is anyone else noticing as of today the new “settings” gear visible when in the references/Get Help/Feedback window? Clicking on it allows the user to change the language and also view connections in a webpage. Not finding any documentation around this being planned change. It is available in Prod\Sandbox for Teams and Web Assistant. See images below.  

I have a compound action that is around 100 lines long but is only 6 actions. I cannot publish or test it as I continue to get the following error:Message: Sorry we cannot process your request this time.I have already broke this set of actions out from another compound action as I wasn’t able to save that one until doing so. I believe this has to do with the size of my compound action as a while back I saw a 5mb size limitation error message when running into this error in another situation. Is my only course of action to break this compound action out into further smaller compound actions?