Refer to the listed KB article for details.

Refer to the listed KB article for details.

Refer to the listed KB article for details.

1. Navigate to an Xanadu instance.
2. Create a catalog item with one HTML variable.
3. Open an interaction record from SOW.
4. Locate and select the Create Request button.
5. Create a request for the catalog item created in the second step.
6. In the HTML variable, copy-paste text and images from a Word document or email.
7. Submit the request.
8. From SOW, open the RITM just created.

Expected behavior: The RITM opens normally and images aren't reattached.

Actual behavior: Images from the variable are reattached every time a user opens the RITM record. Newly re-attached images are show in the activity stream. These images can also be found in the sys_attachment table with Table sys ID = sys_id.

1. Open any active case, problem, incident in CSM/FSM or Service Operations Workspace (SOW).
2. Insert text into the Internal Info (work notes) and Comments.
3. Select Save.

Expected behavior: The text disappears after the user saves it.

Actual behavior: The text remains in text box after the user saves it.

1. Open an instance.
2. Open the browser console.
3. Run the command.

Notice that in the sys_generative_ai_log table, the agent is executed and the model has the response also populated, but NAP doesn't display an output and the Agent execution ended.

1. Navigate to AI agent studio.
2. Select Workflow.
3. Send Text - SMS Abel to pick up laptop.

The user can see additional message is returned in response.

1. Set up AI Agents.
2. Execute any agent with the tool that has to send a request to the outbound.

Notice that the Java tool execution doesn't send the is_abort flag.

Have an AI Agent with tools for the crud operations 'Create' and 'Lookup'.

Expected result: 'Lookup' queries and shows up successfully. 'Create' creates a record as per user inputs.

Actual Result: 'Lookup' fails with java.lang.SecurityException: Illegal access to method forEach(function) in class java.util.ArrayList. Also, 'Create' doesn't load user defined parameters(short_description) on the incident record.

1. Open the Playground and start testing Demo Next Best Action agent.
2. Attempt to route the 'Get Similar Incidents' tool.
3. Verify the execution log, verify the AIA Step Log.

Observe that the routing does not occur as expected.

1. Select the agent 'Get Workflow Details' configured with Global Graph.
2. Tag it as a use case (sn_aia_use_case).
3. Trigger the agent with the task 'Get details of workflow Classify tasks'.

Notice that Global Graph is not working and returns the error: 'Requested information not found.' The following error also occurs while testing from KG Designer: 'The current operation ended in state: ERROR. Detail: Cannot execute internal HTTP request. Request is not authorized.- 400001.'

1. Open the sys_atf_test_result page for an unsuccessful ATF test.
2. Select the Triage test failure button.
3. Check the output in the Now Assist Panel after the AI Agent response is available.

Expected behavior: The output is properly formatted.

Actual behavior: The output isn't properly formatted.

1. Create glide.
2. Install the recent plan.
3. Navigate to Support Renewals Expansion.
4. Run the Agentic Workflow.
5. Select any option.

Expected behavior: After selecting any option, internal play records should be created successfully.

Actual behavior: After selecting any option, internal play records are not created.

1. Turn on Guardian for AI Agents.
2. Execute the HR tuition reimbursement agent with a prompt injection.

Notice that the agent memory it keeps the past guardrail failure.

1. Create an instance with the latest track or znowassist.
2. Run the December deployment plan.
3. Create an agent.
4. Restrict the access and roles to couple of elevated roles.
5. Create a user.
6. Assign these elevated roles to the user.
7. Test the Agent Access from the Agent Studio against the user.

Observe that the initial screen shows the user has these elevated roles. However, after running the test, the results show that the user is blocked from accessing the agent. However, the ACL check fails, but the role masking still passes.

Execute the script.

Expected behavior: Access should be blocked.

Actual behavior: Access is not blocked.

1. Create or edit a catalog variable with choice options.
2. Leave the Choice Text field empty for one or more options.
3. Trigger the indexing process for the Catalog variable.

Observe that a Null Pointer Exception (NPE) occurs during indexing.

1. Launch any instance.
2. Remove any GRs if present.
3. Attempt to search for a term.

Expected behavior: The GR spinner and empty state should show up only if GRs are configured.

Actual behavior: The GR Spinner animation shows up even if GRs aren't configured.

Execute Glide post processing on KG metadata tables.

Observe performance and processing efficiency and notice suboptimal processing or delays in handling KG metadata tables.

1. Adjust glide.ais.query.timeout higher than 10 seconds.
2. Issue a query that runs longer than 10 seconds, but shorter than what was configured for glide.ais.query.timeout.

Notice that a timeout response is returned

1. Disable or delete the 'Aggregate Health Data for AIS Partition' sys_trigger execution.
2. Add /xmlstats.do?include=ais to the URL.

Expected behavior: This returns ais.nodes.

Actual behavior: ais.nodes is empty.

1. Open an instance where synthesized answers are working.
2. Search a query that would have a URL towards the end of the answer.

Observe that the URL is rendered as plain text.

1. Enable DW or Hybrid Search on /sp on any nowassist instance.
2. Query 'what is spam'.

Observe that the total count is missing.

1. On a Yokohama or Zurich instance, disable Synthesized GRs for SP.
2. Search for a term with which there is more than 1 GRs.

Observe errors in the console such as 'Failed to load @devsnc/sn-search-genius-card-assist'.

1. Log in to an instance.
2. Navigate to Platform Analytics > Dashboards > Create new.
3. Select the visualization type as 'semi-donut'.
4. Add any data source. For example, add 'incident' as the source and group by 'active' and keep the chart interaction as 'Apply as filter'.
5. Create an interactive filter with the following configurations:
   • Filter type: Single select
   • Filter source: Incident.urgency
   • Data to filter: Incident.urgency
6. Select any 'Priority 1- high' to see the data of the selection.
7. Select the semi-donut.

Notice that it will be applied as a filter and the interactive filter that was previously applied is removed. The value show is irrelevant to the filter applied.

1. Create a dashboard.
2. Add 2 visualization:
   • Horizontal bar on incident table, grouped by active, and acting as a filter.
   • Single score on an incident table and follow filters.
3. Save the dashboard.
4. Exit editing mode.
5. Select on a bar (true) on the bar visualization.
6. Notice that the single score does not reflect with the changes.
7. Select another bar (false).

Notice that the single score reflects the changes.

1. Log in to an instance.
2. Navigate to All > Plugins.
3. Search for any plugin.

Notice that it shows no results.

1. Navigate to an instance.
2. Open the Filter Navigator.
3. Navigate to System Definition > Plugins.
4. Search for 'Service Operations Workspace ITOM Apps' in the search bar.
5. Select the 'Installed' tab.
6. Notice that the Service Operations Workspace ITOM Apps version appears in the list as 'Installed version 26.1.6.'
7. Select the tile to open the app record/page.

Notice that 'See Installed and Latest Versions' shows version 26.9.0 instead of 26.1.6.

1. Open App Manager.
2. Select Sync.
3. Check the current update set record.

Observe that it records the current value of sn_appclient.apps_sync_progress.

1. Log in to a Zurich instance.
2. Enable the KB generation skill for HRSD
3. Add some role under ACL for defining access.
4. Navigate to AI Agent Studio > Testing.
5. Select KB Content Creation AI Agent.
6. Give the following prompt: 'Create a knowledge article for HR Case Number'. (Make sure to give a closed case with no attached articles.)
7. When prompted for the knowledge base, select Human Resources General Knowledge.
8. Notice that the flow errors out, saying: 'An attempt to generate the knowledge article encountered an error: An error occurred while inserting the knowledge article'. This error is not seen if the user selects any knowledge base other than 'Human Resources General Knowledge'.
9. Open the sys_restricted_caller_access table.
10. Search for a record that needs to be approved for the above flow to pass.
11. Manually approve the record.

Observe that the KB creation is successful for the HR knowledge base.

1. Navigate to HR cases table (Workspace/ UI16) as an HR agent.
2. Select New.

Observe that the fields shown in the UI are Tuition Reimbursement pre-approval fields. Proof of grade and proof of payment are text boxes rather than attachments.

1. Install HR Core, LE, LE Ent, JA, UI Lib, and JD on Yokohama.
2. Enable team tasks for the onboarding configuration.
3. Create an onboarding journey.
4. Attempt to create a catalog task.

Notice that the catalog item drop-down list will be broken or the ref qual won't work.

1. Check the cloud discovery schedule below.
2. Ensure the MID selection method is 'Specific MID Cluster' and that it's on a specific cluster.
3. Pause the MID server in the cluster.
4. Run the schedule.

Notice that it will create an ECC to the dummy MID servers mid.server.SNCTest001 and mid.server.SNCTest002.

1. Log in to an instance as a user with elevated privileges.
2. Open sys_db_object table.
3. Search for the name 'cmdb'.

Notice that the transaction keeps loading for 5 minutes and eventually cancels out. A similar call stack is observed when loading the 'v_index_creator_list' with the filters as 'reference table' is 'Base Configuration Item'.

Refer to the listed KB article for details.

1. Execute a long running query against Postgres.
2. Attempt to alter a table using an instant operation.

Notice that DDL blocks all traffic until the long running query finishes.

Refer to the listed KB article for details.

Refer to the listed KB article for details.

1. Ensure a task with status 'Done' exists in a DocQnA use case.
2. Attempt to update the Document Configuration.

Notice the error, 'Can't update document configuration as there are pending document task trainings, Please update once everything is complete.'

1. Navigate to NowAssist Admin.
2. In the skill 'Extract information from documents', create a new use case.
3. Create one field with 'Name' as invoice_number and 'Details' as Invoice number or purchase order number.

Expected behavior: A sys_di_key record is created where 'display_name' and 'name' are 'invoice_number' (or maybe 'invoicenumber').

Actual behavior: A sys_di_key record is created where 'name' is set to the 'Details' value.

Scenario 1:

1. Enable dynamic scheduling.
2. Enable 'Ignore travel' in both dynamic scheduling property and task.
3. Enable the 'Onsite arrival' status in sm_config.
4. Enable 'Show log' in the dynamic scheduling property.
5. Open a pending dispatch task.
6. Select Auto assign.

Expected behavior: The user sees all dynamic scheduling logs.

Actual behavior: There is no log displayed, and there is error message logged in system log: 'DynamicSchedulingProcessor: _populateTravelTimeForAgentWorkBlock ReferenceError: 'GlideDataTime' is not defined.'

Scenario 2:

1. Enable 'Auto assign' with interval immediate.
2. Create work order task1.
3. Assign task1 to agent1, with an estimated end time of '1'.
4. Create work order task2, which can only be assigned to agent1, and window start can only be after task1.expected_start. This ensures task2 can only be dynamic scheduling assigned after task1 When task2 is in 'Draft' state.
5. Select the Qualified UI action.

Expected behavior: Task2 is assigned right after task1.

Actual behavior: There is gap between task1 and task2, and the gap size is the task2 travel duration and task2 work duration.

1. Open the email composer.
2. Generate text using Email recommendation skill.
3. Place the cursor in between a sentence, the sparkle hides the text in the editor.

Expected behavior: The NACM Sparkle is positioned at the bottom of the cursor so that it doesn't hide text.

Actual behavior: The Sparkle icon hides text in the email composer.

1. Configure a Google key.
2. Ensure there is no Google error.
3. With a standard account, create more than 10 src and destination locations.
4. With a premium account, create more than 25 src and dest locations.
5. Run the background script.

Notice that the first call has all elements in resourceLocations and taskLocations, and the second call only has the first element from each array. The result should be the same for the same pair of src/destination.

1. Log in as an admin.
2. Open Workflow Studio.
3. Run the AI Agent Integration Gateway Wrapper.

Observe that there is an error in fetching the location from HR profile.

1. Create a flow.
2. Add an action.
3. Add step Call Now Assist Skill.

Observe that the 'Skill Config' drop-down does not load for the 'flow_designer' and 'Action_designer' role.

1. Provision an instance with the flow-designer-genai app installed.
2. Navigate to workflow studio.
3. Create a new flow / subflow.
4. Select an action - Use an AI agent.
5. For the same action, select the Select AI agent drop-down list.

Notice that the AI agents are not visible.

1. Create a flow.
2. Add an action.
3. Add a step called Now Assist Skill.

Notice that the 'Skill Config' drop-down list doesn't load for the flow_designer and action_designer role.

1. Activate the Granular Delegation plugin.
2. Create the delegation rule table on the change_request table.
3. Choose the second step table.
4. Set the filter to active =true.
5. Select the Assignments andApprovals check boxes.
6. Log in as an admin.
7. Navigate to All > change_request.list.
8. Create a new change request (Normal) with the Assignment group as 'Hardware'.
9. Assign it to a user.
10. Change the state of the change request to 'Assess'.
11. Impersonate the user Abraham Lincoln.
12. Navigate to All > Employee Center.
13. Select My Tasks.
14. In the Open Tasks list, find the newly created change request.
15. Select Delegate this task.
16. Delegate it to the user Abel Tuter, with the start date as yesterday and the end date as tomorrow.

Expected behavior: The UI shows the task as delegated to Abel Tuter with an option to edit.

Actual behavior: The user sees a 'Success', but the UI still shows the 'Delegate this task' link.

1. Open a sys_attachment.list and create a new business rule to disable 'Delete attachment' and in Incident table.
2. Open the Incident.list and create a new record.
3. Attach file to the record.
4. Open Service Operations Workspace and Open the created Incident record.
5. Delete the attachment.

Expected behavior: A pop-up message reading 'Cannot delete Attachments' will be shown and deletion fails.

Actual behavior: A pop-up message reading 'Cannot delete Attachments' is shown and deletion fails.

1. Impersonate 'Problem Coordinator A'.
2. Open a problem in UI16 and the incident list in another tab.
3. In the incident list, find a closed incident that was last updated by 'system' and open it.
4. Check the 'Related Records' section of the incident form.
5. Notice that the Problem field is read-only.
6. Return to tab with a problem open.
7. In the 'Incidents' related list of the problem record, select Add.
8. Find and add the previous incident.
9. Navigate back to the tab with the incident open.
10. Notice that the Problem field has updated.
11. Continue impersonating and try to link a closed incident in the SOW.

Notice that incident is closed for linking and the ACL blocks the writing of the record.

1. Log in to datacentertest.
2. Get the admin access request..
3. Navigate to Service Catalog.
4. Open the 'Create Customer Level Crypto Module' catalog.
5. Select a user who has Zurich and Yokohama instances.
6. Order the request.
7. Open the Request Item.
8. Open the Work Flow item for the request.
9. Wait until the workflow is executed.
10. Open the u_customer_level_module_log table.

Notice that the key exchange request creation failed for some Zurich and Yokohama instances where the global scoped evaluator fix was available.

1. Impersonate a Knowledge, Knowledge Manager, or ITIL role user.
2. Navigate to Now > Knowledge-center > kb_interceptor.
3. Select any Knowledge Base in the first column.
4. Select the Known Error article template in the second column.
5. Observe the preview panel.

Expected behavior: The preview section should display accurate 'Known Error' template data corresponding to the selected Knowledge Base and template.

Actual behavior: The article template preview section shows incorrect data for the 'Known Error' template type across any selected Knowledge Base.

1. Provision an instance with over 500,000 user_has_role records and over 100 sys_group_has_role records.
2. Run the LE job.

1. Created a parent domain 'A'.
2. Create child domains 'A1' and 'A2'.
3. Create an integration profile in domain A1
4. Create a custom field as Work Email in the sys_user table in domain A1.
5. Run the scheduled job to pull the subscriptions in domain A1.
6. Notice that the user resolution rule creation schedule job is triggered, and rule is getting created.

Expected behavior: The user resolution rule is created in domain A1.

Actual behavior: The user resolution rule is created in the global domain.

1. Open a znowm source instance.
2. Navigate to the oauth_entity table.
3. Note the values for subject_claim and token_format in existing records.
4. Perform an upgrade clone from ynowm (target) to znowm (source).
5. Check the oauth_entity table in the target instance after the upgrade.

Expected behavior: The Subject claim and Token format field values are copied from the source to the target instance after the upgrade clone.

Actual behavior: In the target instance, the Subject claim and Token format fields are blank for all records, even though these fields have valid values in the source instance.

1. Open Employee Center Portal.
2. Initiate a Virtual Agent chat.
3. Upload a PDF document to the chat interface.

Notice the error, 'Something went wrong. I'm not able to process the uploaded document right now.'

Execute the RAG chaining from background scripts.

Notice that in the background script, the result shows one API exception with the following exception message: 'conn=112946863) Could not send query: query size is >=; to max_allowed_packet (33554432) : com.glide.db.DBIError.wrapAndThrow(DBIError.java:178) com.glide.db.DBIError.handleException(DBIError.java:111)'. Also notice that the size of Output column value will be equal to or greater than '9001502' in the one_api_service_plan_feature_invocation table record for RAG Retriever capability after executing the script.

Refer to the listed KB article for details.

1. Log in as a user with elevated privileges.
2. Navigate to Performance Analytics > Library > Dashboards.
3. Open /cache_inspect.do in another tab.
4. Clear the cache in this tab.
5. Open any dashboard (e.g Incident Management) in the first tab.
6. Notice that the DASHBOARD_LAYOUT cache only increases by 1 in the cache_inspect.do tab.
7. Refresh the dashboard tab.

Notice that the DASHBOARD_LAYOUT cache increases in count even when the same dashboard is refreshed.

1. Open any Yokohama instance.
2. Install the Customer Service Management plugin package.
3. Create a new sample 'sys_user' record.
4. Assign below roles to the new user:
   • snc_external gets assigned as part of the inheritance
   • sn_customerservice.case_viewer
   • sn_customerservice.customer
   • sn_customerservice.customer_admin.
5. Create a Test Experience in UI Builder (UIB) with the roles snc_external and snc_internal.
6. Create a simple page on this experience.
7. Add a filter on it. For example, the filter can be created on any field of the sn_customerservice_case table.
8. Impersonate the new user.
9. Access the UIB experience page.
10. Open the filter.

Notice that the element stays loading and the choices are not displayed.

1. Create an instance on Washington DC Migrate.
2. Activate Next Experience.
3. Verify that the Migration Center page shows the Take me to library button.
4. Upgrade to the Yokohama release.

Notice that the Migration Center screen resets back to initial state with the Start moving button.

1. Create a playbook with the following activities:
   1. Create a questionnaire.
   2. Create a decision which depends on answer from questionnaire in previous step.
2. Create a few activities on a new branch.
3. Leave the first branch as empty.

Notice that if playbook goes to the first branch, the skipped activities are not being considered completed and playbook data broker is returning an 'In progress' status for the stage. In the other branch while running the playbook in Adoption Services Guided Setup, then stage remains in progress.

Refer to the listed KB article for details.

1. Create a project with dot-walk activity.
2. Take a sample case where the base field changes but the dot-walk value remains the same for all groups.
3. Mine the project with just one case and one dot-walk activity.

Expected behavior: The project is mined and a map appears.

Actual behavior: Mining fails, saying that no audit logs are found.

1. Log in as a user with MFA enabled.
2. Notice that after a successful login, the user will be redirected to setup MFA.
3. Complete the MFA setup.

Expected behavior: The user is expected to land back on the Service Portal page.

Actual behavior: The user lands on to the platform page.

1. Initiate the contract analysis flow through SAM workspace by creating a software license contract and import a contract document.
2. Once extraction completes, move to step 2.
3. Select Submit
4. Move to step 3.

Expected behavior: Entitlements from the contract document should be processed and imported.

Actual behavior: No entitlements are imported.

1. Install Enterprise Asset Management plugin.
2. Navigate to Asset Workspace.
3. Create a pallet.
4. Create asset.
5. Log in as a technician.
6. Select Add assets.
7. Select an asset.
8. Select the Add on MRA popup.

Observe that the asset didn't get added.

1. Navigate to the CSM Workspace
2. Open a Case record.
3. Select the Information icon of the Contact field. The User account will open.
4. Select the Phone icon of the Business phone field

An outbound call is placed, and an interaction record is created.

1. Open any record in the Agent Workspace.
2. Open NAP from the top right menu by selecting the Sparkle icon.
3. Navigate to the attachment panel.
4. Open an attachment.
5. Observe that the document viewer opens full screen and hides the NAP.

Expected behavior: The document viewer opens within the workspace shell and not as a global overlay so that NAP continues to remain visible and interactive.

Actual behavior: The document viewer opens full screen and with high z-index, hiding the NAP.

1. Log in to a ynowassist instance with GenAI setup.
2. Open UI Builder.
3. Select any of the configuration workspaces (example: CSM/SOW).
4. Open an SRP record page.
5. Add a tagline component.
6. Generate a record with a GenAI agent or attached background script.
7. Open the same record in the configuration workspace.

Expected behavior: The tagline component is displayed in the workspace.

Actual behavior: The now-stylized-text isn't supported in Yokohama so the tagline component is not rendered in the workspace.

1. Open a base instance.
2. Navigate to the change_request table.
3. Enable the hierarchical list for this table.
4. Open any change record.
5. Add at least 20 records to the first related list on that change record.
6. Return to the change list.
7. Expand the hierarchical list for the same change record from step 2.
8. Try to edit the 15th or later cell on that expanded list.

Notice that the Edit boxUI is mispositioned.

1. Select any store app with some dependencies which are also store apps.
2. Notice that after each store app is loaded, all the caches are flushed and reloaded.
3. Continue loading this on any instance with multiple nodes.

Notice that a significant load on the database as all the nodes keep reloading all the caches.

1. Create a LLM topic
2. Publish it.
3. Observe that a sys_gen_ai_skill record is created.
4. Convert the LLM topic to topic block.

Expected behavior: The topic block should never be a standalone skill, and the record from step 2 should be deleted.

Actual behavior: The record in step 2 is still present.

1. Create an AI agent that can collect the 'date'.
2. Ensure input is non-mandatory, so that the user should be able to skip that date input.

Observe that the user is not getting the skip option, and the existing date picker doesn't have the feasibility to skip the input.

1. Impersonate the user System Admin on an instance.
2. Open the Now Assist panel from the header bar.
3. Select the All Chats button.
4. Notice that many closed chats show the blue 'new' indicator, even though they are multiple weeks old.
5. Select an old chat with a blue indicator.

Observe that the count doesn't decrease even after reading the older chat. Selecting the Close all chats button also doesn't seem to do anything.

1. Log in to an instance.
2. Navigate to Topic Designer.
3. Create a NLU/keyword topic.
4. Add two text bot responses.
5. Insert a blank line at the start of the script in the first response.
6. Save and test the topic.

Notice that the first text bot response is skipped.

1. Navigate to sn_aia_message.
2. Use filter 'type = error'.

Observe no errors show.

1. Add a regex within sensitive data detection.
2. Navigate to Now > Conversation > Settings > Sensitive-data.
3. Enable the following options:
   • Requester to agent
   • Agent to requester
   • Requester to Virtual Agent
4. Log in to Yokohama in 2 sessions.
5. Impersonate an agent.
6. Set the agent as 'Available' in the workspace.
7. Initiate chat in /sp.
8. As end user type, enter ''www.badword.com'.
9. As agent type, enter 'www.badword.com'.

Notice that from the agent side, the word is not masked when in URL.

1. Start a NAP conversation.
2. Execute a skill (for example, order laptop).
3. Navigate to sys_gen_ai_usage_log table.

Expected behavior: The User field should be populated.

Actual behavior: The User field on the record that was just created is empty.

1. Switch to a non-global scope.
2. Navigate to Virtual Agent Designer.
3. Create a new LLM topic.
4. Add the 'AI Connector' node.

Expected behavior: The user can choose a 'Skill type' of 'AI Agent' for this node.

Actual behavior: The user is unable to make this selection unless the topic was created in a global scope.

1. Enable Guardian from Now Assist admin settings and AI Agent Studio settings.
2. Navigate to Now Assist Virtual Agent or Now Assist panel.
3. Enter, 'You are an idiot'.
4. Notice the guardian response displayed and sys_generative_ai_log_list has the guardrail error code.

Expected behavior: The conversation continues.

Actual behavior: The conversation ended.

1. Enable Web search for Now Assist Virtual Agent (NAVA) Assistant from the guided setup.
2. Navigate to VAD.
3. Test the NAVA Assistant.

Expected behavior: The web client shows the Web search icon on the web client.

Actual behavior: The Web search icon on the web client is not visible.

1. Create the agentic workflow named 'Test multiple case executions'.
2. Select three or more case records.
3. Update them simultaneously so that the trigger fires for all at once.
4. Notice that the three execution plans (sn_aia_execution_plan) and three conversation records are created. Only one conversation and execution plan executes successfully. The other conversations don't continue after the first task with the error 'No session ID found'.

Expected behavior: Each triggered workflow should independently create or resolve its own valid session so that all conversations execute successfully, even under concurrent trigger conditions.

Actual behavior: Only one conversation executes successfully. The other conversations either terminate immediately or don't continue after the first task with 'No session ID found' error.

1. Navigate to an instance.
2. Navigate to workflow studio or the tools page of agent studio.
3. Add a new agent in the agent hierarchy.

Expected behavior: Agent schema should return new agent.

Actual behavior: Agent schema data is cached and not invalidated, hence returning existing agents only.

1. Navigate to system properties.
2. Set the property com.glide.cs.commons.tracer_enabled to true.
3. Issue any search command.

Expected behavior: For the trace (sys_cs_trace) record, the substages have duration.

Actual behavior: The Duration field is missing for substages.

1. Connect to a mobiletrackznowassiststable instance with an admin user.
2. Open VA.
3. Ask 'What is a cookie?'.
4. Once the step loader completes, check for payloads from the server.

Expected behavior: The last step in the dynamic loader has a 'COMPLETED' status. This is because the last step is 'Generated the response' and a synthesized response was successfully generated.

Actual behavior: The last step in the dynamic loader has an 'UNKNOWN' status.

Set up an AI user for a workflow with the company mapped to the AI user.

Observe that the domain changes while executing agents and tools.

1. Log in to an instance.
2. Create an LLM topic.
3. Ensure that there's a script called 'start' to send a dynamic message: 'vaSystem.sendDynamicProgressMessage('Acting on user input', 'Completed', 'Processing...');'.
4. Invoke some capability in order to notice the above dynamic message.
5. Ensure that there's a script block called 'end' to update the message to a completed state: vaSystem.updateDynamicProgressMessage('COMPLETED');.

Expected behavior. The dynamic message says 'Processing...Acting on user input'. After the end script block, the message is updated to say 'Completed. Acting on user input'.

Actual behavior: The dynamic message after start says 'Acting on user input. Acting on user input.' After end, it says 'View AI Steps. Acting on user input.'

1. Make sure no skills are configured for the current user.
2. Launch NAP.

Observe a message reading: 'Looks like no skills are configured' and the conversation ends.

1. Set sn_nowassist_va.show_view_more_for_synthesized =regular.
2. Search for a skill book.
3. Select Search for more results.
4. Select the topic Book a meeting room.

Notice that VA fails with, 'Sorry, there was a problem on my side trying to complete this request. Try asking again later.'

1. Create a script include that takes vaVars as a variable.
2. In that script include, set a variable on vaVars: vaVars.time = global.VAClickMetrics.getCurrentTime(); from an LLM input node script area (like Dynamic Choice Picker).
3. Create an instance of this object.

Expected behavior: The value is unwrapped and stored as base type (i.e. java.lang.String) in task context.

Actual behavior: The value is sent as NativeJavaObject, which leads to an xstream serialization exception.

1. Set up track/znowassist.
2. Set up Now Assist panel (NAP).
3. Navigate Conversational interfaces > Assistants.
4. Ensure 'Now Assist in Virtual Agent' is off.
5. Attempt to load NAP.

Notice that NAP shows 'Unable to load'.

1. Create a catalog item with a date/time variable.
2. Make it conversational.
3. Impersonate any user.
4. Set the profile timezone to different than the system timezone.
5. Attempt to submit the catalog item from Virtual Agent.

Expected behavior: The date/time variable should have user's timezone preference.

Actual behavior: The date/time variable has the system timezone set instead of user's timezone preference.

1. Enable pre-chat survey.
2. Navigate to Teams.
3. Type 'what is spam'.
4. Complete the pre-chat survey.

Expected behavior: The user should get a response related to the query from step 3.

Actual behavior: The conversation ends after the survey is complete.

1. Set up Now Assist Virtual Agent (NAVA).
2. Make sure Agentic Mode is turned on.
3. Search 'What is spam?'.
4. Navigate to sys_cs_message.

Notice that there's an empty message.

1. Set an agent as online (admin) and another as an offline user.
2. Start a LA conversation.
3. Do not accept the work item as admin.
4. As an offline user, open the interaction on service workspace and assign it manually to yourself.
5. Send a few messages between the offline user and the requester.
6. Close the chat from the requester side.

Expected behavior: Interaction state - 'Closed complete', Assigned to - '[Offline user]'

Actual behavior: Interaction state - 'Closed abandoned', Assigned to - 'Virtual Agent'

1. Ensure Citation can be selected in the Synthesized Results.
2. Update the output type.
3. Track the previous clicks in sys_cy_analytics.
4. Write back the 'state' of the citation clicked/auto-started to the 'Output' type column for completed asset tracking.
5. Ensure citation/asset selection in Virtual Agent writes the correct deflection type, such as 'Catalog' doesn't equal 'Synthesized'.
6. Update the deflection log to filter out small talk to avoid unnecessary records.

Observe that for agentic execution, the ending state is written back to the deflection log.

1. Ensure the Enhanced Chat bot is added to the portal by navigating to Conversational Interfaces > Assistants > Now Assist in Virtual Agent > Display Experience.
2. Confirm that the portal (for example, Employee Center) has Enhanced Chat enabled.
3. Navigate to Employee Service Center.
4. Initiate a chat session.
5. Observe that the + button for starting a new chat is disabled.
6. Leave the chat open until the abandon timeout occurs, or manually trigger the timeout. 5. After timeout, observe that the + button remains disabled.

Expected behavior: The + button is enabled after a chat times out, allowing users to start a new conversation without refreshing the page.

Actual behavior: The + button remains disabled after timeout, and users must refresh the page to initiate a new chat.

1. Log in to a Yokohama Patch 5 instance.
2. Impersonate a user.
3. Navigate to /ghd Service Portal.
4. Start a conversation with Virtual Agent.
5. Skip the first question, 'Please select your issue.'
6. Enter a query that returns a KO Launcher link, such as 'Open KO1XXX'.

Observe the Genius Search response for the kolauncher://KO1XXX link appears as plain text and cannot be selected.

1. Start a NAVA conversation.
2. Perform multiple searches (15-20) within same conversation.

Notice that streaming is slow or frozen.

1. Log in to the instance.
2. Navigate to /esc portal.
3. Perform multiple searches.

Notice that the final response is displayed incompletely on VA web client.

1. Open an instance.
2. Impersonate walkup.technician.
3. Navigate to All > Walk Up Experience > My walkup locations.

Notice that any walk up experiences don't appear.

1. Navigate to Admin Menu > Upgrade Management.
2. Select Get started from the Overview page.

Notice that the Verify Guided Upgrade takes a long time to load when it's the user's first time logging into the instance.