Agent Studio

Hello,I have a cmdb lookup skill which returns the business application list, I am been trying to instruct the LLM to display only name, description and application owner. tried below map code also tested on data mapper playground, but getting response saying no records. without below code, it does display all the results. when there is single result, i would like to show something like below, but no much information on logs and no failures also bot is saying just ,,”No ownership information was found in the business application database or internal knowledge sources.”please help me to identify the issue. also give me a good documentation back end supported syntaxes like system_prompt, display_instructions, I want to study deep here.  {  "result": [    {      "short_description"

The recent Agent Studio Updates & Next Steps email mentions Agent Studio (Classic) will be retiring in the H1 2026. Will Events be a part of this retirement or will they move to Agent Studio? If so, are Webhooks the intended replacement for Events?We use Events in many messaging campaigns and would like to plan for the future. 

Is it possible to customize ticket resolved ticket notification to include a link to a survey? Or what is the recommended way to build this out using Agent Studio to serve up the survey?

My team is working on a Databricks SQL statement and wanted to see if anyone may have any insights. It looks like our bot isn't showing the responses when we test. The logs show that the actions are running without errors, but the conversation logs show that the output is coming through as null. Since the mapping is Moveworks-specific and a bit complex, I’ve attached the logs so others can take a closer look and help troubleshoot. Statement is something like this, I’ve redacted proprietary information:{  "statement": "SELECT DISTINCT <InvoiceID>, <SupplierName>, CONCAT(<BusinessUnit_PO>, ' ', <PurchaseOrder>) AS PO_ID, <VoucherStatus>, CASE WHEN <VoucherStatus> = 'STATUS1' THEN <PaymentDate> ELSE NULL END AS PaymentDate, CASE WHEN <VoucherStatus> = 'STATUS1' THEN <PaymentID> ELSE NULL END AS PaymentID, CASE WHEN <VoucherStatus> = 'STATUS2' THEN <ScheduledPayDate> ELSE NULL END AS ScheduledPayDate, CA

I’m working on an Outlook use case where we need to fetch meetings for a user within a specific start and end datetime using the Microsoft Graph API.ChallengeWhen calling the CalendarView endpoint, the Graph API requires both start and end datetime values. If no time offset is included, it defaults to UTC.For example, for Eastern Standard Time (EST), the offset should be UTC-5 hours.The question is: How can we dynamically pass the correct offset based on the user’s timezone?What we’ve triedWe attempted to format the datetime using the Moveworks DSL Format_Time function, but this requires the timezone in IANA format. From the Graph API, we only get the user’s timezone in Windows format (e.g., Eastern Standard Time). Is there a recommended way to convert Wind

If I setup a connection using authorization code, it will always make the actions based upon the initiating user. With this setup, I need the ability to have Moveworks make actions using someone else’s access. Here are some examples where this is needed:The initiating user requests for something that requires approval. A moveworks approval process is used to manage that approval. When the approver approves it, I need that to initiate an action in the same system but now using the approver’s account as the initiating user does not have access to approve their own request. With the new listener functionality, I see no way to run any action associated to a connection that has the authorization code grant setup as there is no initiating user.

Do we have the option to reuse plugins that we have developed in Agent Studio? We’ve created an approval plugin for one team within our organization, and now another team is requesting a similar solution. Although each team has different data points and systems they connect to, they both want the same approval flow process.I’m trying to determine if it’s possible to reuse the existing plugin to avoid hitting our plugin limit and to streamline development. If there’s a way to customize or clone the plugin for different data sources while maintaining the core process, that would be ideal.Could you confirm if this is feasible or suggest the best approach to efficiently handle multiple teams with similar workflow needs? Thanks!

We’re bringing together developers, innovators, and AI enthusiasts for a full-day hackathon at the iconic Chase Center in San Francisco on Tuesday, October 7, 2025. This one-of-a-kind opportunity to build right on the court puts you at the center of innovation where champions compete.✨ Special for Moveworks customers: Bring 2 or more developers from your org, and you’ll be entered for a chance to win a pair of the cutting-edge Meta AI Sunglasses*: smart glasses that bring AI right into your everyday life.Whether you’re building your first AI agent or pushi

I have a list of questions and answers in a restricted google document. It is in a sheet, but it can be moved to a doc as sheets do not seem to be supported. The use case is to ask the bot many different questions that are included in the document and have the bot return with the exact answer for each one of them. Currently this mostly works, but the answers are getting summarized. The answers need to be 100% exact as they are approved internally as correct. If I tell the bot to return the exact answer, it works most of the time, but is there a way to have it always return the exact answer instead of having to ask every time?

Is there a workaround for the output mapping result limit? I'm working on a flow where a user can query the bot to return all of their tickets in JIRA, but I'm losing some data in the process when the mapping gets bigger. Using conversational plugin, not compound action.Example Output Mappings:Mapping 1MAP(): converter: key: item.key project_name: item.fields.project.name status: item.fields.status.statusCategory.name summary: item.fields.summary url: $CONCAT(["https://company.atlassian.net/browse/", item.key]) items: response.issuesreturns:1. Project A – 38 tickets2. Project B – 1 ticket3. Project C – 1 ticket4. Project D – 13 tickets5. Project E – 32 ticketsMapping 2MAP(): converter: assignee: item.fields.assignee.displayN

Hi All,I am looking at removing a button presented in a notification to the user.Specifically I want to remove a button so the user only has two options. The button I want to remove from the notification shown is “Edit request” to avoid user confusion.Currently selecting “Yes” instructs the middleware to execute a remote action against the machine and clicking “Yes” again will fail as the remote action cannot be executed more than once on the middleware system.I cannot see anything in the CREST Event workspace that would refer to what looks like a consent option to proceed with the choice as shown above.