Agent Studio

Hi all, We have our prod environment set up with Moveworks configured to have a Live Agent Handoff with Glia. We want to make sure our Sandbox is configured properly too. We have a site ID that we inputted as well as a queue ID, but we’re unsure of what our Webhook Token should be - if it’s generated already, how to generate it if it isn’t, or how to retrieve it if it already exists. Do we need to setup/configure a connector? Right now when a live agent handoff attempt is made in our Sandbox environment, the live agent interaction/request is not seen on Glia’s agent side so it seems that it is not reaching Glia at all, and the response on our Moveworks bot is “I attempted again to start a live agent chat for you using the description you provided; the request first went into a pending state and then failed with the internal error LIVE_AGENT_UNABLE_TO_START_CONVERSATION, meaning the system was unable to open a conversation with a live agent.”

We have built a use case where user can requests monitors by submitting tickets in ServiceNow directly via MW bot.Required slots:Requested_for - User data typeMonitor_type - static resolverrequested_date - string with date validation policy (checking for 4 business days atleast DSL →  $PARSE_TIME(value, "%Y-%m-%d") >= $TIME().$ADD_DATE(0, 0, IF $INTEGER($TIME().$FORMAT_TIME("%w")) IN [0, 1] THEN 4 ELSE (IF $INTEGER($TIME().$FORMAT_TIME("%w")) == 6 THEN 5 ELSE 6))delivery_address - String - validation policy DSL →value.$LENGTH() >= 15 AND value.$MATCH("[0-9]").$LENGTH() > 0 AND value.$MATCH("[a-zA-Z]").$LENGTH() > 0additional_info: String Conversational process flow is below → Decision policy (mandatory monitor type slot) → if X monitor is selected

Hi team,I’m trying to enforce a 1MB file‑size limit on file uploads captured through a Moveworks slot. I attempted this using a Script Action, but it appears the approach won’t work because:The location field of an uploaded file is not a valid URL that can be fetched. Script Actions run in a sandbox and cannot directly access Moveworks internal file storage. Because of this, the script cannot read the file bytes or calculate file size, so validation never triggers.I wanted to check whether Moveworks currently provides any supported method or API to validate file size before sending the file to a destination system via HTTP Action.Is there an officially recommended pattern or upcoming feature for this?Thanks in advance for any guidance

Hey agent builders!Python Actions dropped on this session and the energy was immediate — An attendee hit us with a "no effing way" in the chat before I even finished the sentence. We also shipped a brand-new docs site, rolled out knowledge and form content activities, and then spent the back half tackling pre-submitted questions and a live DSL debugging session that humbled me in real time.Ryan and Steve were both on as co-hosts. Solid crew, solid session. Here's what went down. New Docs Site — docs.moveworks.com We shipped a completely rebuilt documentation site. The old sidebar situation is gone — now there's a product switcher that breaks things into AI Assistant, Enterprise Search, Agent Studio, IT Service Management, and the API Reference. The search is actually useful now: keyword results plus an AI assistant tab that can answer follow-up ques

If you've built anything non-trivial with APIthon, you've hit the ceiling. Complex data transformations, real calculations, text processing — either impossible or painfully hacky in a restricted scripting subset. The most consistent ask from builders has been: "I need real Python with real libraries."Done. Python Script Actions let you write full Python 3 directly in Agent Studio — with access to all built-in modules and a curated set of pre-approved external libraries. They plug into Compound Actions and Conversational Processes exactly like APIthon scripts, but with dramatically more power. Define input arguments from slots or previous step outputs, write your logic, and the value of the last line becomes your output. That's it. What you get:→ Full Python 3 with all standard libraries (json, datetime, re, math, collections, etc.)

If you found an ambient agent on the Marketplace, you already know the pain: there's no install button. You're staring at someone else's fully-built agent and manually recreating compound actions, trigger configs, webhook listeners, and mappings by hand. Nobody was doing that.That's fixed. You can now install ambient agents directly from the Marketplace into your environment in a single flow. Webhook plugins, scheduled trigger plugins — they all install cleanly with all their compound actions and mappings imported automatically. Same process also powers dev-to-prod promotion, so sandbox-to-production workflows are first-class now.What's included:→ One-click Marketplace install — ambient agents now have an Install button. No more manual recreation.→ Automatic asset import — compound actions, trigger configurations, and mappings come along for the ride.→ Smart asset resolution

We want to start building a use case to let users view the required skills for their requested job profile. We have a couple of questions to assess the feasibility for a client. Here are the requirements for reference: Enable employees to do the below via the bot: View required skills for their job profile and  View their most recent skills assessment (self + manager)  View their entire history of skills assessment  View required skills for any given job profile  Show a delta of skills and skill levels between any two job profiles for example, show me the Skills required and the expected skill level for Level 7 Design Release Engineer vs Level 6 Design Release Engineer or between my job profile and Level 7 Designer  Manager can ask and view skills for any employee in t

Is it possible to reference and manipulate a user attribute within a Content Activity? For example: grab a user’s email username@email.com and remove the @email.com so that it returns only username

Depending on the response to a button in my compound action I need to direct a user to a website, with some parameters attached.I currently wait for them to press the button then display the link but this is an additional step that isn’t really necessary.It would be more efficient if they could click the button and be taken directly to the web page.Is this possible?

I wanted to flag a potential scalability observation, and keen to have further discussions.With the introduction of Ambient Agents, more systems are integrating with Moveworks via webhooks. As we continue to add plugins across multiple systems, the volume and frequency of inbound calls will increase significantly, which raises the risk of hitting rate limits (HTTP 429 errors).To illustrate with a concrete example:One Power Automate flow pulls data from Intune and sends device status alerts to Moveworks users. A second Power Automate flow also pulls data from Intune and sends device status alerts to Moveworks users. A ServiceNow flow sends SAM reclamation alerts to Moveworks users.If these plugins are scheduled to run daily, the cumulative execution time becomes quite long. While these jobs are running, additional webhook calls from external systems can still occur, increasing the likelihood of rate limiting.We

In my application, all request bodies need to be form data with data as the key and the json body as the value. I am able to handle this in http actions with raw json body by having it be something like data={<insert body here>} and having the header of Content-Type=application/x-www-form-urlencodedI want to use the new data mapper functionality in http actions so it will handle escaping and such for me, however I cannot seem to find a way to get it to work with this setup. I have attempted numerous times with variations of data: <json to mapper syntax> , but the app always returns with a missing data element error. Is there any way to handle this currently?

Brief summary: New tickets often lack context agents need to troubleshoot quickly. This pattern uses a Moveworks Ambient Agent to automatically pull relevant enterprise data at ticket creation time, run LLM actions to correlate + summarize data, and post a single Agent Brief to the ticket before first touch, reducing repeat investigation and accelerating MTTR. IssueNew IT tickets usually arrive with limited context. The agent spends the initial minutes doing repetitive investigation:What changed recently? Has this happened before? Who owns the service? What systems/dependencies might be involved? What’s the best next step?SolutionAutomatically generate one consolidated Agent Brief at ticket creati

You can now collect files from users right inside Moveworks! With File Slots, you can design plugins that request and receive files — like receipts, documents, and invoices — and send them directly to external systems.  What's included:• File uploads via a built-in "Upload File" button in chat• 🎕 New: Multiple file uploads now supported via List[File]• Seamless integration with HTTP Actions through flexible file mapping• Supports raw upload, multipart/form, multipart/related, application/octet-stream, and base64 encodingSupported file types: .png, .jpg, .jpeg, .pdf, .txt, .csv, .xlsx, .xlsm, .xls, .xltx, .xltm (up to 5MB each)📖 Learn more:• File Slots docs:

I’m facing an issue with a Compound Action where the flow executes successfully until fetching the manager_record, but the notify step does not trigger. There are no errors in the logs, both employee_record and manager_record are not NULL, and the workflow does not explicitly fail or terminate. The execution simply stops after fetching the manager record, and I don’t see any logs beyond that point. For testing, I’m using a single static record via script, but in the actual use case, the data comes from an HTTP action, which is why a for loop is used. I’m also attaching a screenshot of the logs for reference.Here is the exact Compound Action configuration:steps: - script: code: | return [{ "LOGON_NAME": "mw_test_user_5@company.com", "EMAIL_ADDRESS": "mw_test_user_5@company.com", "MANAGER": "manager-name", "ACCOUNT_EXPIRY_DATE": "2026-02-26 0

I am using a listener to take in requestIds that start an automated process. When setting it up, there were some situations where the requestId was sent twice and I wanted to avoid the processing being called twice, so I added in Deduplication configuration in the listener. The request body being sent to Moveworks is always in this format:{“requestId” : “<id>”}Therefore for the deduplication, I added requestId as the Deduplication key (no spaces or quotes) with a 1 min window. However I am noticing that it is blocking requestIds that are not equal but are sent within the 1 minute timeframe. The difference between them in these cases is the last character in the requestId (ex ABC1234 vs ABC1235) In the logs, I am seeing one go through and “debounced_update” log entries for the other requestIds sent. I have no event filtering enabled for the 1 plugin that is associated to this listener. Please assist me in understanding the root issue and how to set this up correctl

Hey everyone, first time posting here, I’ve tried out the User Consent Auth / Authorization Code Grant between Moveworks and Workday, functionally it is working well after user consented, but is there anyway to detect/test if a user consent expired/non-exists at the beginning of a plugin?Problem StatementIf a user never consent previously or if token already expired, the first time calling a plugin will always fail (not yet consented / token expired). Immediately afterwards, User Consent Request will appear: User can then proceed to consent, but the plugin already ended so user will have to manually retrigger the plugin after consented, and retype everything again. QuestionI need to make my plugin:1) Check if user consent exist/valid2) If false in 1) then show User

Since from two days many users are reporting a popup issue, there is no change made from our end like manifest file for anything w.r.t our bot. its not specific to certain skill, happening for all. when yo cancel it or “Not now”this shows “this card action is disabled because of you blocked the bot” anybody facing similar issue?  

Hi,Is there any workaround to catalog item form fillability for ServiceNow catalog item forms.most of our form are still not fillability due to client scripts.Regards,Rajeev

We have several Events configured in the old Creator Studio (old) that send a Teams notification as soon as the Power Automate flow receives a payload. Two additional events are also tied to this flow. When users receive the notification, they are presented with Approve and Deny buttons, which trigger corresponding Power Automate flows. These flows interact with the Planview API to patch the required task numbers.However, an issue has been observed:If a request is created (for example, on 30th January) and the user attempts to approve or deny it after 3–4 days, the event‑based path no longer triggers, causing the downstream process to stop.We would like to know whether there is any platform limitation or time‑bound restriction where a static Teams message (triggered by an event) can only remain actionable for a certain duration,

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

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 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 succ