Agent Studio

Hello Community!I wanted to share a set of practical use cases for managing device health that empowers end-users to take action when their hardware needs attention. While these notifications and workflows leverage Nexthink, other device management tools can be substituted using a similar architecture.The Solution: Nexthink Ambient AgentsNexthink offers a Remote Actions API to initiate tasks, native webhooks for real-time alerts, and Nexthink Query Language (NQL) to define data triggers. Combining these capabilities allows ambient agents to listen, notify, and assist users with actions such as:Clear Device/Teams Cache Restart Device

We redesigned the Developer Hub into the single page I wish existed when I was learning Agent Studio. Quick links to basically everything you need: docs, quickstarts, cookbooks, marketplace, API reference, and more.Everything you already use is still there (Quickstart Guides, Cookbooks, Purple Chat Builder, Academy, extensibility docs). Just easier to find.The goal was simple: one place, no hunting. Building a plugin, troubleshooting a slot resolver, figuring out where to start. Developer hub gets you there in one click.Go poke around and bookmark it: https://www.moveworks.com/us/en/developers you might see a familiar face

Encountering the following issues with Agent Studio Quickstart Guide Exercises: 1) Quickstart #3: Basic Process Agent: Not able to generate/view any Purchase Requests in Moveworks Purple API as shown below in the screenshot. Tried creating multiple session IDs, but this step won’t work. Link to exercise: https://docs.moveworks.com/agent-studio/quickstart-guide/quickstart-basic-process-agent  2) Quickstart #4: Structured Data Analysis (SDA): Facing issues with the “Fetch Accounts” endpoint and receiving the following 401 error:Link to exerc

Has anyone successfully setup the Docusign Connector and installable plugins from the Marketplace? We loaded the installable plugin and believe we have correctly setup the connector using Oauth2 and JWT Auth.The trouble is hitting when we try and run the HTTP Action: Docusign_Get_Account_Id_By_User_Email. I keep getting a 401 error. Confirmed with Docusign support that the consent is setup correctly, but unsure what’s missing. Thought it might be the authorization in the header not being passed, but again unsure as this is our first go at this Plugin. 

Hi Team,We are developing a use case to change the profile picture. This is a pretty straightforward use case where we have a File data type slot and an API call to upload the photo to Workday.Workday does have a couple of validations for an uploaded photo that we are trying to check via Moveworks:1. Check whether the uploaded picture is an image (jpeg, png) - we are doing this by checking the file name in the validation of the slot.2. Check the file size. (probably less than 1MB)3. Check if the image is a square or a certain aspect ratio.4. Can we validate the optical recognition - if the uploaded picture is a passport, a driver's license, etc.?I asked the Agent Architect bot, but the recommendation it gave wasn't possible (a script action for the file size).

I have a notify step that uses RENDER() with a generated input string training_str. The template uses HTML tags (e.g. <b>) and a training_str built from a $MAP / $CONCAT expression, but the newline characters render literally (you see \n) when the message is sent to Teams. I’ve also tried "<br>" and "<br >" and they also end up displayed literally. How can I generate a string from my inputs that will render new lines appropriately in Teams? MS Teams screenshot of how the message renders with a literal \n Notify YAML (relevant parts only): - notify: output_key: notification recipient_id: data.user_mw_record.user.record_id

Hey builders,Steve dropped the Structured Data Analysis GA announcement, then the session turned into a resolver strategy clinic with invoice lookups, custom object slots, and an LLM-powered ticket routing architecture. Ryan co-hosted, Steve did the heavy lifting.What we covered:Structured Data Analysis (SDA) Now GA — code interpreter for large datasets, auto-triggers at 7K tokens Resolver Input Arguments — fixing dynamic resolvers that return the entire table Custom Object Slots — collecting structured data like attendee email + required/optional flag Output Mappers — stripping raw API noise so the reasoner shows only what users need LLM Actions for Ticket Routing — full webhook-to-assignment architecture with generate structured value New Agentic AI Docs — LLM fundamentals, context windows, and

I have a workflow where a Compound Action presents a button to the user. Clicking this button triggers a Conversational Process via a Data Mapper payload. Inside this process, the bot asks the user for an input (e.g., a description or amount_to_add slot) and eventually reaches an HTTP action with a Consent card.The Issue:If the user decides to cancel at the consent step (by clicking a Cancel button or typing “cancel”), the flow ends. However, if they scroll up and click the original button to restart the process, the bot remembers their previous answer due to session caching and skips asking the question again.What I’ve Tried:To force the bot to start fresh, I tried passing an empty value for the slot directly from the trigger’s Data Mapper, but ran into blockers:Passing null unquoted resulted in an “Invalid Data Mapper” YAML schema error. Passing “null”

I noticed a thing, why some plugins have the option to ‘Install’ while in others I have the option to ‘Build’ ‘Get Notified’?     

Hi,I received the attached “Page Not Found” error indicating I need access to the latest version of Agent Studio to complete my Moveworks Agent Studio Developer: Level 1 Certification.Could you please guide me with the next steps to get access.Thanks,Charlene

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.”

I’ve setup a Rapid Global notification when a user signs in or out with Webhooks.The challenge I have is I don’t want to enable the notification for everyone, however, Webhooks doesn’t allow you to select an audience to restrict, the Audience option is disabled.It would be useful to be able to control the notification audience on the Moveworks end rather than have to adjust 100s of users in Rapid Global.  

Hello folks, I’m new in Moveworks however I’m familiar with the plugins, Studio developing, and Setup ingestion/connectors. My organization is looking to integrate Kronos with Moveworks, and I was wondering what’s the standard workflow after setting up the UKG plugin in Moveworks, do I need to set up an API Controller in Kronos to enable communication, are there specified endpoints?

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 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

I am getting an error while trying to test a compound action which states  Message: Sorry we cannot process your request this time.Error Code: 13121Trace ID: YTkNgH59WP66ivn2What does this error means ? How to get over this

Hi everyone,We’re currently building Ambient Agent use cases in Agent Studio and are running into the same challenge mentioned in this thread — there isn’t a way today to access Ambient Agent analytics.While we understand from Moveworks that proactive notification analytics are on the roadmap or that we can use the Data API (we don’t have it setup yet), we’re trying to understand what other customers are doing right now to monitor success and adoption for Ambient Agent use cases in the meantime.For those of you who have Ambient Agents in production:Are you capturing any success or usage metrics? If so, how are you doing it? Using notification logs? Integrating with downstream systems (ex: ticketing platforms, Power Bi dashboard,

We can find the available time of our colleagues using the findMeeting API and this will give us the available slots of the user. This works perfectly for my calendar.Usecase: For the executive assistant, to get meetings on behalf of a shared or delegate mailbox that they have access to, this API doesn’t work.During our implementation:The findMeeting API works correctly for the authenticated user:https://graph.microsoft.com/v1.0/me/findMeetingHowever, when attempting to retrieve available meeting times for a shared or delegate mailbox using:https://graph.microsoft.com/v1.0/users/{sharedMailboxEmailId}/findMeetingthe request returns an "Unsupported Scenario" error.Even

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