Public Product Forums

Hello, We are facing an issue with the MW generic approval template(mw.create_generic_approval_request). We are able to build the approval request successfully, but when we trigger it to the user, the notification contains the header as approval_details. Whatever we pass as “approval_details” appears as the header. This is fine when we add one or two lines of approval details, but when we add a list of information for approval, it looks bad for the end user and it’s duplication valuesIs there any way to eliminate or truncate these details in the header?  

Sharing an important update for everyone using the Snowflake Dynamic Query / Snowflake Cortex Analyst plugin.Snowflake has announced that they are deprecating single-factor password sign-ins, including for service accounts. Once this enforcement begins, password-only authentication will stop working, and Snowflake will require MFA or non-password mechanisms for programmatic access.Reference:🔗 https://docs.snowflake.com/en/user-guide/security-mfa-rollout  Our Snowflake connectors today are set up using username + password. If these connectors remain unchanged, they may start failing with authentication errors when Snowflake completes this rollout. Snowflake recommends migrating to one of the supported authentication methods for automated integrations:Key-pair authentication (public/private key)

I’m working on a use case to ingest a Google Drive that requires profile boosting. I see in this documentation clear guidelines for how to label and thereby boost geographic specific information based on user profile location.  Does this hold true for both MIME types of content, articles and files? Context for what an article vs file is taken from this documentation:Knowledge Base Articles (KBA): HTML-content such as .aspx pages & .md/.mdx (markdown) Files: Such as .pdf, .ppt, .doc, & .txt 

Strengthen your data governance with custom viewsToday, we’re introducing a new way to manage who can see what inside your analytics app. This release combines enhanced RBAC controls with flexible, customisable views—giving super admins more control over sensitive data and enabling teams to share only the insights that matter to the right stakeholders. 🔐 Govern Sensitive Interactions With the Interactions Viewer RoleUser–bot conversations can include sensitive information. Previously, anyone with Bot Analytics Admin or Bot Analytics User roles could see all interactions with full PII.With this release, super admins now have a dedicated way to govern that access through a new role: Interactions Viewer.Only users given this role by a super admin can view the full interactions table—including user prompts, bot responses, email, ID, and citation details, etc. This ensures that organisat

How do I keep the bot from returning the resolved input arguments in a reply back to the user? I have tried adding the instruction to “NOT show resolved input arguments in the message to the user if information needed to be resolved” in the plugin description.  Haven’t had any luck with this though in Agent Studio. This was not created via a compound action, but does involve a decision policy with 2 actions. The required slot for the decision policy is is_department. Any help would be appreciated! 

Our organization is redesigning a support portal and we would like to pass user queries from the page’s search to AI assistant for web.We’re fine with redirecting to the AI assistant web page, but we want to link directly into a conversation with the user’s query as the first prompt of a new chat. I can see that chats create a new URL with the following pattern: /assistant/t/<chat-id>, and it stands to reason that the landing page at /assistant is passing the text I enter to the assistant and generating the URL with the chat id.What is the URL where I can pass text directly to the AI assistant to open a page with the assistant already working on the query? Would it be something like /assistant/q?text=’How do I access the VPN’

Hi community,I am having an issue where the llm instructions and guidance are being shown to the user? “Need to call tool again with chat_history. No need to ask user per tool instr; we should summarize ourselves.” “Need to provide chat_history back to tool. They told: first ask user for missing args, then after they provide, call tool before messaging—but system also says don't message before tool except greetings. But here tool explicitly requires ask user. So ask: briefly summarize history ourselves and send as chat_history without needing user? But instructions say ask user and then call tool after user provides. So I must now ask user something, then when they respond, call tool. For now I should ask them to confirm / describe issue and mention reinstall option will come later. Question only.”Does anyone has this issue also and any workaround - I think this is related with slot resolvers.

Hi All, I am trying to add an approval workflow using the inbuilt Action “mw.create_generic_approval_request” however I am confused where to add the approver email address?. I am getting only two input parameters as given below. approval_details: <!-- Placeholder for approval_details -->users_requested_for: <!-- Placeholder for users_requested_for → Any help on this topic will be appreciated.

Hello Moveworks Community! 🚀 We have updated our Interactions API to now include free-text user comments from CSAT campaigns. This gives you direct visibility into the exact feedback and sentiment your end users are sharing. How to retrieve this data:  You can capture the user's specific feedback in detail.content field.To filter for these specific events, look for: Interaction Type: INTERACTION_TYPE_UIFORM_SUBMISSION Label: mw_form Detail Field: "CSAT Feedback"  

You can now collect multiple values for slots from users and pass conversation context directly into dynamic resolvers in Agent Studio. This is enabled for all customersWith List[t](Multiple-Value Slots), you can let users provide several entries at once like multiple users, strings, IDs, or any supported data type without repeated prompts. Collect multiple values for a single slot Works across all supported slot types (String, User, Email, etc.) Returns a clean array for use in resolvers and actions With Dynamic Resolver Context Passing, you can pass data gathered in conversation straight into your dynamic resolvers. Pass slot values, action outputs, and metadata into resolvers Chain resolvers together to create hierarchical slot resolution No more manual workaround

We’re excited to announce two new features that make Enterprise Search easier to use.1. Date Modified Filter 🗓️We’ve added a new filter that lets you narrow search results by last modified date, using both pre-set options and custom date ranges. This makes it easy to focus on the most recently updated content, helping you find the latest information faster and avoid outdated results. 2. Rich Markdown in Summaries  📝We’ve added enhanced formatting in Enterprise Search summaries to make results more structured and readable. Summaries may now include headers, bullet points, code lines, code blocks and bolded details for a clearer, more organized presentation of information. These visual improvements make it easier to scan key points, understand structure and extract insights quickly.

For Answers/knowledge search, sometimes the bot returns a response with links/URLs directly in the response, but sometimes it provides just text and user has to dig into the references to find the links. Is there a way to force the response to include the URLs so that user can navigate to information more easily?Also, is there any way to update the GenAI prompt that’s fed to the bot so that we can have the bot respond more appropriately according to our organization’s needs?

I’m wondering if any other customers are awaiting their November 2025 Knowledge Studio recommendations. Our recommendations are typically received by the 3rd of each month and as of 11/14/2025 they have not been received. I’ve been unable to confirm if this is specific to our account or KS as a whole through support. Please Help!

hey all, i have an Event notification setup that i have a need to pull some log data out of for the last 2 months of the Event plugin running. i’ve looked in the Analytics modules though, and it says my Events have no data to report for the timeframe…i also checked the Logs module and tried to do some filtering but also came up empty handed :/ my use case is essentially: there is a job that runs in our ITSM and at the end of it sends a payload to the Event endpoint to deliver to a recipient(s), i need to retrieve all the recipients that the message has been delivered to over the last two months.i noted these below two community posts from a few months ago and it seems like it’s known that this capability isn’t supported currently, but i wanted to triple check and confirm there wasn’t another workaround to accomplish my goalt

We have many articles in our platform that contain a list of dates. These could be holiday dates for a specific year(s), pay dates for specific year(s), etc.  When users ask for the upcoming date(s) in the specific list, the bot often serves incorrect dates compared to the prompt and current date.  Example:Prompt: When is the next holiday for US Employees? Bot response: The upcoming holidays for US employees are New Years Day January 1, 2025.  You’ll see the year is incorrect. We also have other instances where employees were served previous months, not upcoming months.  Any ideas for optimizing the content for better serving? Also interested if this could be looked into as a potential improvement for the Moveworks platform to further optimize this type of interaction. 

Hi Team,I’ve recently noticed that Knowledge and Forms features in Agent Studio Content Activity seem to be unstable for me. Specifically,When I add a knowledge article in a content activity, the bot displays empty “” results. For forms, instead of  the usual buttons to open them, I only see a hyperlink.Is this a bug? Has the product team already identified this issue? Thanks in advance.

We’ve made it even easier to find the information you need. Search results will now group related documents, snippets and chat messages together, so you can see all the most relevant information in one clean, organized view. ⚡ Find answers faster: Related content appears next to each other, helping you browse results quickly and focus on what matters. 🗂 Stay organized: Multiple versions or iterations of the same document are now combined into a single, tidy group. 🔭 See the full picture: Snippets from the same document and messages from the same thread are neatly grouped for easier understanding at a glance. This update makes your search experience cleaner, faster and more intuitive, helping you get to the right information with less effort.Examples:

Hi All,I am working on a simple compound action use case where I am pulling data from one of the ServiceNow tables. Even though the use case is quite straightforward, it doesn't seem to be working as expected.This is one example, but I have observed similar behavior with other use cases as well.I requested fields like number, justification, and status. I have also mentioned these fields in the use case description, but the result is not aligned with the code written. Compound actionResult Anyone experienced similar behavior with Agent Studio.  Regards,Rajeev

Hey Developers!Last week we kicked off our very first Developer Office Hours and wow – the energy was 🔥 many of you showed up live, we crushed a bunch of real-world questions you submitted ahead of time, and we even squeezed in a live demo of the brand-new File Slots feature that just dropped into Limited Preview.Here were the hottest topics we tackled together: Slot descriptions & smart defaults → How to make the assistant infer “today” vs “tomorrow” timestamps without always prompting the user (hint: clever slot description + decision policy + DSL date functions = 🤝) OAuth 2.0 user consent flows (Google Workspace, GitHub, etc.) → Live walkthrough of the authorization-code grant + the exact connector settings you need so your agents can act on behalf of real end-users Dynamic slot resolvers & output mappers →

Is there a way to customize the bot response or use the decision policy to route the response based on error status code received from an API call? Example:I have an action that submits an email. If the email is valid the API returns the required data if not valid it returns Status Code 400 error. Status Code: 400Response:{ "code": "error", "message": "Email not valid."}Email not found in the system means their account doesn’t exist and the bot should relay access request information. I’ve attempted using the decision policy using if data is valid or not but with no luck.

You can now collect files from users right inside Moveworks!With File Slots, you can design plugins that request and send files (like receipts, ID documents, invoices.) directly to external systems.Single-file uploads via a built-in “Upload File” button in chat Integration with HTTP actions through flexible file mapping Supports raw upload, multipart/form, multipart/related, application/octet-stream, and base64Want to learn more about the feature (& see a demo)? Check out our latest docs here:https://help.moveworks.com/docs/file-slots#/ https://help.moveworks.com/docs/http-action-files#/ Limited PreviewWe’re launching File Slots in a l

I am trying to build a conversational process that can target either a user's email or computer name. Is it possible to configure a slot to automatically handle user input as follows?If the input is an email address, use it as the slot value. If the input is not in an email address format, attempt to find a user by that name and use the matching user's email address as the slot value. If no matching user is found, use the original input text as the slot value.I plan to use a decision policy to process this slot differently based on whether it’s an email address or a computer name.Thank you.

I read somewhere that Moveworks supports OBO workflows for Confluence but can’t seem to find the docs for it. Can someone point me to the documentation for this? 

We are looking to open up the ability to users to drag and drop documents into the chat bot so that they can be sent to other destinations.  What we are finding (Chat Bot is in Teams) is that when someone drops a file in chat, it puts that file in the users own OneDrive but the chatbot only knows about the URL and Name of the document not and of the Data in the document.Has anyone else attempted this and how have you done it?  If not is this something that may be coming in the future?I feel like this could be really userful even if it was something like how Brief Me does it and the document is stored for X amount of time then removed.  But atleast you could still process the documents.  Im thinking something like HR documents, Claims Documents, etc.

I noticed that in the user attribute documentation there is a list of default values that can be mapped like first name, last name and so on. When i look at a user profile or the user lookup settings there are additional attributes that are not mapped and show empty on the profile like (Preferred Name, seat name, seat link, etc.).  I created a custom attribute for seat name and it populated as a second attribute called seat name so in the user profile now I have seat name filled as a custom and seat name empty separately. Are these able to be mapped without creating a custom attribute?