Fixed (Patch Release): Service bridge name change in Agentic AI to Service Exchange.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Updating our dependent app versions for better packaging.

Fixed: Fixed an issue with mismatched API contracts for Intent Activity Analysis (Process Mining).

Fixed - ACLs for voice input being sent to NowLLM.

Read only fields will be now "strict" read only in general Now Assist skill tables. With this change tables cannot be updated by using client scripts, browser consoles et al.

Fixed: Fixed dependency issue that caused upgrade issues on Yokohama patch 11 (YP11).

Removed: Spoke Generation has been removed from the Now Assist for Creator Bundle and will be a separately installed app on the ServiceNow Store

Fixed (Patch Release): Service bridge name change in genai to Service Exchange.

Removing Now Assist for SOMT dependency.

Fixed (Patch Release): Service bridge name change in Agentic AI to Service Exchange.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Updating our dependent app versions for better packaging.

Fixed: Fixed an issue with mismatched API contracts for Intent Activity Analysis (Process Mining).

Fixed - ACLs for voice input being sent to NowLLM.

Read only fields will be now "strict" read only in general Now Assist skill tables. With this change tables cannot be updated by using client scripts, browser consoles et al.

Fixed: Fixed dependency issue that caused upgrade issues on Yokohama patch 11 (YP11).

Removed: Spoke Generation has been removed from the Now Assist for Creator Bundle and will be a separately installed app on the ServiceNow Store

Fixed (Patch Release): Service bridge name change in genai to Service Exchange.

Removing Now Assist for SOMT dependency.

In this version, the Generate hardware asset insights generative AI skill aggregates data from related records to produce comprehensive asset analysis summaries - covering the asset lifecycle, chain of custody, audit status, and financial metrics to support key asset management activities.

Fixed (Patch Release): Service bridge name change in Agentic AI to Service Exchange.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Updating our dependent app versions for better packaging.

Fixed: Fixed an issue with mismatched API contracts for Intent Activity Analysis (Process Mining).

• The Knowledge Center is a centralized hub for Knowledge Admin and Manager that is an entry point for KM Management personas to manage all capabilities and configurations.
• Guides users on the next-best actions to take by displaying actionable key insights.
• Is a suite of KM Tools for suggesting content creation, updates, duplicates, and article retirement.

Fixed - ACLs for voice input being sent to NowLLM.

Read only fields will be now "strict" read only in general Now Assist skill tables. With this change tables cannot be updated by using client scripts, browser consoles et al.

Fixed: Fixed dependency issue that caused upgrade issues on Yokohama patch 11 (YP11).

Contract entitlement data extraction skill as part of Now Assist for Software Asset (SAM).

Removed: Spoke Generation has been removed from the Now Assist for Creator Bundle and will be a separately installed app on the ServiceNow Store.

Updates for YP13 compatibility. This section outlines key improvements and adjustments made to ensure seamless integration with the YP13 system, including protocol enhancements and configuration updates.

Fixed (Patch Release): Service bridge name change in genai to Service Exchange.

• Resolve SaaS connection errors - This feature enables diagnosing and resolving connection errors occurring in SaaS integrations within ServiceNow SAM, either at the validate connection step during initial setup or during the normal integration flow. Investigate root causes such as authentication failures, API endpoint changes, or credential expiry. It helps in providing actionable error details at the point of failure:
  • Whether during setup or live sync.
  • Reducing time spent on manual troubleshooting.
• Contract entitlement data extraction - Reduce manual effort by leveraging AI to automatically create entitlements in SAM by extracting licensing information from software contracts.

Removing Now Assist for SOMT dependency.

YP13 compatible release.

Fixed: Fixed automatic execution for new instances.

Fixed (Patch Release): Service bridge name change in Agentic AI to Service Exchange.

Complaint Case AI Agents: Update fields tool allows arbitrary field modification.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Updating our dependent app versions for better packaging.

Fixed: Fixed an issue with mismatched API contracts for Intent Activity Analysis (Process Mining).

Fixed: ACLs for voice input being sent to NowLLM.

Read only fields will be now "strict" read only in general Now Assist skill tables. With this change tables cannot be updated by using client scripts, browser consoles et al.

No Features delivered via this application, it is a wrapper for the Complaint Case AI Agent collection.

Removed: Spoke Generation has been removed from the Now Assist for Creator Bundle and will be a separately installed app on the ServiceNow Store.

Fixed (Patch Release): Service bridge name change in genai to Service Exchange.

Fixed: Fixed automatic execution for new instances.

Fixed (Patch Release): Service bridge name change in Agentic AI to Service Exchange.

Complaint Case AI Agents: Update fields tool allows arbitrary field modification.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Added support for reasoning levels for NowLLM.

Updating our dependent app versions for better packaging.

Fixed: Fixed an issue with mismatched API contracts for Intent Activity Analysis (Process Mining).

Fixed: ACLs for voice input being sent to NowLLM.

Read only fields will be now "strict" read only in general Now Assist skill tables. With this change tables cannot be updated by using client scripts, browser consoles et al.

Added a new skill that generates approval checklist.

No Features delivered via this application, it is a wrapper for the Complaint Case AI Agent collection.

Removed: Spoke Generation has been removed from the Now Assist for Creator Bundle and will be a separately installed app on the ServiceNow Store.

Fixed (Patch Release): Service bridge name change in genai to Service Exchange.

• Added
  • Mobile Authentication: OAuth token support for secure mobile access.
  • Configuration Management: Scripted API for mobile voice service configuration.
  • Call Source Differentiation: Origin-aware processing for mobile, phone, and browser.
  • Fallback Strategies: Channel-specific fallback handling.
  • Analytics Tracking: Device type and origin tracking for reporting.
• Fixed
  • Minor security and functional fixes.

• Changed
  • RAG Popular Searches is now used when integrated chat experience is enabled.
• Fixed
  • Corrected the Trigger Condition for Search Personalization Rule in guided setup.
  • Addressed an issue that prevented the AIS Next Experience search application record from being found when enabling NAP.
  • Fixed an issue where regular search results were not appearing when NAP was enabled for global search.
  • Resolved an issue where selecting "Use existing configuration" in Virtual Agent during profile setup did not correctly copy the configuration.

Complaint Case AI Agents: Update fields tool allows arbitrary field modification.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Fixed: Set the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only. This configuration change enables seamless functionality for Advanced tier customers.

Fixed: Updated the prompt to resolve a tool call issue in the regulatory analysis agent.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only, supporting seamless functionality for customers on the Advanced tier.

Fixed - ACLs for voice input being sent to NowLLM.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the IRM Advanced tier.

• New
  • General Availability - Agent Miner graduates from Innovation Labs to GA.
  • Auto-generated evaluation datasets - Every opportunity now ships with an evaluation set for validating agent performance.
  • Opportunities list as default view - The list page is now the primary entry point for reviewing automation opportunities.
• Changed
  • Custom agent creation restricted to Prime SKU and current Now Assist packages - Foundation and Advanced tiers no longer support custom agent creation.
  • UI enhancements across the product based on customer feedback - easier navigation, clearer opportunity details, and refined settings.
• Fixed
  • Numerous UI and setup issues resolved for GA.
• Removed
  • Duplicate-and-customize existing agents - removed based on customer feedback.

• Added support for exploration of data from secure scopes.
• Enhanced query generation capabilities such as:
  • Support for global OR conditions.
  • Improved filtering on class field of CMDB tables (support for InstanceOf operator).
  • Added ability to filter by sys_created_by/sys_updated_by fields.
  • Semantic segment tuning improvements (ability to disable specific segment types).
• Administration enhancements:
  • Added module for AI Data Explorer logs (includes error and performance logs).
  • Added module for workspace configuration for AI Data Explorer.

Fixed: Inconsistencies within plugin.xml and pom files.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Fixed: Set the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only. This configuration change enables seamless functionality for Advanced tier customers.

New: Added support for Agent Advisor.

Fixed: Updated the prompt to resolve a tool call issue in the regulatory analysis agent.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only, supporting seamless functionality for customers on the Advanced tier.

Fixed: ACLs for voice input being sent to NowLLM.

• Changed: Updated Endpoint for Create Response Actions - The endpoint for Create Response actions now includes the version segment (v1) in the resource path, aligning with the latest Azure OpenAI API specification.
• Fixed: Data Type Format Constraint Issue for Spoke Vision Wrapper - Fixed a data type format constraint issue in the spoke vision wrapper where a field typed as object required input in a strict format - any deviation caused failures or incorrect processing.

Agentic workforce UI enhancements.

• UI size enlarged in UI16 when Polaris is switched off.
• Change summarization is displayed for a few change requests, but not for others, after the upgrade. Now Assist for ITSM to 28.7.0.
• NACM accessibility issues with header labels.

Contract entitlement data extraction skill as part of Now Assist for Software Asset (SAM).

Fixed: Minor defects.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the IRM Advanced tier.

Track which knowledge articles and catalog items support successful virtual agent deflections instead of transferring to human agents using the ITSM Virtual Agent Analytics dashboard.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the Advanced tier.

• Improved Sources and Citations Experience - The way users access sources and supporting evidence has been redesigned to be more intuitive and consistent across the platform:
  • Replaces the previous "info (i)" icon with a new interaction pattern aligned to platform standards.
  • Makes sources and citations easier to discover and review.
  • Aligns with emerging design patterns to reduce cognitive overhead.
• Stronger Consistency Across the Platform - The updated citations experience follows the same interaction model used across other AI-powered features, creating a more unified and scalable design approach. This consistency improves usability and reinforces trust in AI outputs.
• Clearer AI Presence and Role Distinction - To support the introduction of ServiceNow's autonomous workforce, the product now more clearly distinguishes between different types of AI activity:
  • New visual indicators differentiate AI Specialists from standard agentic workflows.
  • Updated avatars and presence indicators make it easy to identify which AI entity is acting within a record.
  • Consistent experience across both workspaces and the classic UI.

• New
  • General Availability - Agent Miner graduates from Innovation Labs to GA.
  • Auto-generated evaluation datasets - Every opportunity now ships with an evaluation set for validating agent performance.
  • Opportunities list as default view - The list page is now the primary entry point for reviewing automation opportunities.
• Changed
  • Custom agent creation restricted to Prime SKU and current Now Assist packages - Foundation and Advanced tiers no longer support custom agent creation.
  • UI enhancements across the product based on customer feedback - easier navigation, clearer opportunity details, and refined settings.
• Fixed
  • Numerous UI and setup issues resolved for GA.
• Removed
  • Duplicate-and-customize existing agents - removed based on customer feedback.

Fixed: Fixes related to incident search.

Updated to support the latest version of the dependent apps.

The ITOM MCP server now exposes the SLO Creator toolset: Incident Linked CI Reliability, CMDB CI Reliability Status, CMDB CI Reliability Topology, and Create SLO for CMDB CI.

• Added support for exploration of data from secure scopes.
• Enhanced query generation capabilities such as:
  • Support for global OR conditions.
  • Improved filtering on class field of CMDB tables (support for InstanceOf operator).
  • Added ability to filter by sys_created_by/sys_updated_by fields.
  • Semantic segment tuning improvements (ability to disable specific segment types).
• Administration enhancements:
  • Added module for AI Data Explorer logs (includes error and performance logs).
  • Added module for workspace configuration for AI Data Explorer.

• New
  • Get a quick overview of the AI Desktop Actions application by using the onboarding wizard that highlights steps related to recording, refining, testing, and activating desktop actions.
  • Use the Show Inputs / Show All buttons in the Test modal to filter required input fields.
• Changed
  • The name of the application is now changed to AI Desktop Actions from Agentic Desktop.

New: Generate an action plan for a record from external documents.

• Skill to support creation of widgets from natural language prompts - Describe what you need in plain English and the Widget Builder generates a fully functional AIXF widget, including component code, server script, input schema, and metadata. No manual coding required to get started.
• Skill to support updation of existing widgets via prompts - Modify any existing widget by describing the change you want. The builder intelligently applies only the necessary updates while preserving all existing functionality, styles, and data integrity.
• Automatic design system and accessibility compliance - Every widget created or modified through the builder automatically adheres to AIXF design patterns, theming tokens, and WCAG 2.1 AA accessibility standards out of the box.
• Built-in input validation and ambiguity handling - The builder validates all inputs before processing and will not modify or generate code when a prompt is unclear, instead responding with specific guidance on what information is needed to proceed.

• New
  • Added support for GPT 5.4, OpenAI's latest large language model, for LEAP automation recommendations and AI-assisted workflows.
• Changed
  • The following default models are updated: claude_large to Claude Sonnet 4.6, claude_small to Claude Haiku 4.5, gpt_small to GPT 5.2 mini.
  • New model: gpt_large to GPT 5.4. Prior versions remain available for customer selection.

New: Introducing the ITOM MCP Server, available through the MCO Console Admin. The ITOM MCP server enables seamless integration with your ServiceNow instance, allowing you to connect to and manage alerts efficiently. This capability extends to leveraging any 3rd-party AI tool, such as Claude, ChatGPT, and others, to retrieve information, perform analysis, investigate, and access additional skills to streamline your workflow.

Fixed issue with navigation to view all records from list visualization response.

Fixed: Inconsistencies within plugin.xml and pom files.

Changed: Added additional recommendations to improve performance for catalog items with the Lookup select box variable type.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

• New
  • GA of Employee Slate conversational experience.
  • Users with the new role sn_na_analytics.ai_engmt_viewer will have the "View" right to the "Analytics" tab within Assistant Designer.
  • Branding configuration for the Premium Chat experience.
• Changed
  • The logic of the Test Panel drop-down menu.
• Fixed
  • AI Agents shouldn't be discoverable to Now Assist Panel - Platform (default) Assistant, unless the customer has contacted support.

Fixed: Inconsistencies within plugin.xml and pom files.

Fixed: Some defect fixes for supporting translation in other languages.

Added support for Australia content from ServiceNow documentation.

• Bi-directional conversational interface with features like filtering tasks, conversational org chart, blended chat-search experience with access to external content connectors.
• Enhanced agentic experiences with widgets, catalog items, and actionable notifications for an interactive and personalized experience.

• Extended InCountry SKU Support - Enabled support for in country routing for AU and IN.
• Enabled new version for Claude and Open AI models, and enabled support for model fallback when requested model is not available.
• Enabled support for Gemini Computer use API for Web agents use case.
• Enabled caching support for Amazon Bedrock models to get latency and cost benefits.

Fixed: Set the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only. This configuration change enables seamless functionality for Advanced tier customers.

Fixed: Inconsistencies within plugin.xml and pom files.

Fixed: Inconsistencies within plugin.xml and pom files.

• Issues with ZTSD triggering flow in autopilot mode.
• Next wave fixes for TR.
• l10n fixes.

New: Added support for Agent Advisor.

Fixed: Updated the prompt to resolve a tool call issue in the regulatory analysis agent.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only, supporting seamless functionality for customers on the Advanced tier.

• Service Mapping now includes two autonomous AI Agents that eliminate the manual effort of reviewing ML-powered candidates and maintaining CSDM relationships at scale.
• The Service Mapping AI Agent for Service Maps creation automatically generates service maps from high-confidence Service Fingerprints and persists the full topology in the CMDB, with created maps set to non-operational for Service Mapping admin review; eliminating the need to manually discover, build, and generate service maps.
• CSDM Business Application to Infrastructure AI Agent automatically connects Business Applications to their matching discovered Application Services using AI semantic search, working top-down through all business applications on a recurring schedule.
• This feature requires: ITOM AI Agents for Service Mapping application (sn_sm_gen_ai, version 1.2.6 or later), Service Mapping Plus version 1.20.0 or later, Zurich Patch 9 or Australia Patch 2.

First release of this app includes a guided onboarding checklist for Discovery and Event Management setup.

• The Knowledge Center is a centralized hub for Knowledge Admin and Manager that is an entry point for KM Management personas to manage all capabilities and configurations.
• Guides users on the next-best actions to take by displaying actionable key insights.
• Is a suite of KM Tools for suggesting content creation, updates, duplicates, and article retirement.

• Uptaking GPT 5.4 for T2R and Description generation.
• Few defect fixes.

Initial release.

Fixed - ACLs for voice input being sent to NowLLM.

• Changed: Updated Endpoint for Create Response Actions - The endpoint for Create Response actions now includes the version segment (v1) in the resource path, aligning with the latest Azure OpenAI API specification.
• Fixed: Data Type Format Constraint Issue for Spoke Vision Wrapper - Fixed a data type format constraint issue in the spoke vision wrapper where a field typed as object required input in a strict format - any deviation caused failures or incorrect processing.

Subflows, Actions, Knowledge Graph, and Scripted REST APIs are now available as MCP Tool types.

Fixed: Email Generation and KB summarization issue.

• Skill details view: Users will now be able to view details about a particular skill in Now Assist Admin Console, Now Assist skills page. By clicking on "view details" for a particular skill, Now Assist admin users can view relevant information about a skill like description, key benefits, and prerequisite skills.
• Archiving skills: Users can now remove skills from the Now Assist skills page and push those into the archival section in Now Assist Admin Console. The archival section can be accessed by clicking on Now Assist skills, archived skills section.
• Model version fallback and non availability: In the model version section users will be informed with alerts about model version when a particular model version is unavailable in a specific region and when a particular model version falls back to a different model version.

• Document search for Approval checklist generation is limited to 3 search results.
• Document search threshold can be customized through the sys_property sn_na_req_agent.approval_assist_agent_document_match_threshold.
• Change in the default model for the Approval checklist generation skill.

Agentic workforce UI enhancements.

New: Added a dashboard for visualizing granular feedback (response feedback) data for Now Assist in Virtual Agent.

• Conversational Administration - Full Settings Coverage
  • Administrators can now manage all global Now Assist Center settings directly through the conversational interface, eliminating the need to navigate separate admin consoles for configuration tasks.
  • Now Assist Guardian controls are fully accessible via the conversational experience, with confirmation prompts before any changes are applied.
• Platform and ITSM Out-of-the-Box Skill Activation
  • Platform-level and IT Service Management out-of-the-box Now Assist skills can now be discovered, evaluated, and activated directly within Now Assist Center.
  • Skill activation flows support both a quick conversational path and a full configuration path for skills that require additional setup.
• AI-Assisted Help for Now Assist Administrators
  • Administrators can now ask questions about Now Assist configuration, Guardian settings, and skill management in natural language and receive answers sourced directly from ServiceNow documentation and training resources.
  • Each response includes references and direct links to source materials, replacing the need to search across dispersed documentation.

• UI size enlarged in UI16 when Polaris is switched off.
• Change summarization is displayed for a few change requests, but not for others, after the upgrade. Now Assist for ITSM to 28.7.0.
• NACM accessibility issues with header labels.

• Scaled the SDG request to accept sample data up to 50 records, enhancing its capacity to handle larger datasets for improved results.
• New OOB datasets to kick start the AI practitioners journey in evaluation for common use cases.

Minor defect fixes.

Fixed: Inconsistencies within plugin.xml and pom files.

• New
  • MCP connectivity: The Search CMDB and Create configuration item agents now expose MCP endpoints, enabling agentic tool integration.
• Changed
  • 3P (third-party) Model support updates.

Contract entitlement data extraction skill as part of Now Assist for Software Asset (SAM).

• New
  • Uptake of Automated quality assurance skill from Service Quality capability.
• Changed
  • Email Reply Recommendation supports case extensions.

Fixed: Minor defects.

Updated to support the latest version of the dependent apps.

Minor defect fixes.

Fixed: Updated orders for new third party models for incident summarization.

No functional changes.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the IRM Advanced tier.

New: Introducing the ITOM MCP Server, available through the MCO Console Admin. The ITOM MCP server enables seamless integration with your ServiceNow instance, allowing you to connect to and manage alerts efficiently. This capability extends to leveraging any 3rd-party AI tool such as Claude, ChatGPT, and others, to retrieve information, perform analysis, investigate, and access additional skills to streamline your workflow.

• New
  • Track which knowledge articles and catalog items support successful virtual agent deflections instead of transferring to human agents using the ITSM Virtual Agent Analytics dashboard.
  • Use the ITSM MCP server to query and retrieve any information within the context of an incident.
• Removed
  • Incident assist skill has been deprecated, moved to the Archive section, and is no longer available for use.

• Changed
  • Support for upgraded versions of third party LLMs has been provided.
  • Enhanced UI for visual clarity in AI-assisted features.
• Fixed
  • The Now Assist summary card header is now displayed consistently across Legal Requests and Legal Matters while switching the records.

Enhanced the prompts for Order Capture and Order Approval Flow for Order Summarization.

Supporting May release versions of Now Assist platform apps.

Updated release dependencies.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the Advanced tier.

• No configuration changes required. All improvements take effect automatically.
• In the March and April releases, the AI team has shipped new third party models for OpenAI and Claude providers. This version uptakes those 3p models.

New: Added support to display Suggested Fixes (Actionable Insights) in AICT as part of production monitoring.

April GA priority defect fixes.

April GA priority defect fixes.

Minor defect fixes to ensure stable operation across supported environments.

• New
  • Portfolio Insights - Identify planning items which are delayed beyond their planned start/end dates auto-categorized by severity with insights into top root causes behind the delays generated by AI. Review recommended actions by AI to immediately take steps to mitigate the impact. Identify planning items with mis-aligned planned/actual dates to approved dates and review recommended actions to correct them.
  • Project Risk Detection - Automatically detect and create project risks using AI to reduce manual effort and improve visibility. AI analyzes project data and generates draft risk records with details such as title, description, impact, probability, AI rationale, and mitigation plan. These are then presented to the project manager for review, allowing them to accept and add them to RIDAC or reject them.
  • Demand manager AI specialist - Validates and converts operational demands autonomously. This feature is currently supported in the controlled availability mode, please contact ServiceNow to implement this specialist.
  • Epic health status: Use Now Assist to instantly assess epic health across six risk dimensions and get a red, yellow, or green status with plain-English reasoning, directly from the epic record page.
• Changed
  • AI Status Reports now computes overall project health (red/yellow/green), generate AI-powered status reports using templates, and produce enhanced executive summaries to create draft status reports for projects.
  • Project Insights now includes additional topics such as budget fluctuations, project delays, projects not started, dependency issues, and critical path changes to provide richer and more actionable project visibility.
  • Goal insights along with AI forecasted status are now generated automatically via a scheduled job (shipped default inactive). AI rationale can be viewed against the AI generated status for each of the goal and target.
  • Enhanced the demand summarization skill to include data from cost plans, monetary and non-monetary benefit plans, demand tasks, resource assignments and work notes to summarize demands.

Minor defect fixes to ensure stable operation across supported environments.

Model creation using Now Assist for OData/BAPI/Tables.

Made minor enhancements.

Made minor enhancements.

• New
  • Agentic AI catalog item requests with enhanced conversational fluidity and performance.
  • Seamless handoff from chat to catalog form for non-conversational catalog items.
  • Available on the Premium chat, a contextual chat experience that appears throughout the platform.
• Fixed
  • Minor enhancements and defect fixes.

Fixed: DocIntel flow failures when Guardian PII masking is enabled. Resolved pre-processing errors caused by invalid characters introduced during PII masking of documents containing emails or addresses.

New NowLLM model Apriel 2.0 and default to latest new 3P providers.

Made minor enhancements.

• Agent Advisor to Agentic Evaluations Integration
  • When evaluating an agent built with Agent Advisor, the Agentic Evaluations module now automatically populates data records, filter conditions, and agent task and context configurations based on clusters created during the advisory process.
• Dataset Selection UX Enhancements for Agentic Evaluations
  • Updated the dataset selection experience within Agentic Evaluations with usability improvements. Easier and more intuitive browsing, selection, and management of evaluation datasets.
• Fallback Model Notifications in Now Assist Skill Kit
  • Skill Kit prompt editor now surfaces clear notifications when a fallback model is used in place of the user's selected model for a given prompt.

• New
  • Analyze Task Trends agentic workflow enhancements
    • Improved actionable recommendations from the trend analysis.
    • Ability to configure input fields for different tables, e.g. incidents.
  • Generate Resolution Plan workflow enhancements
    • New option to run the workflow in fully autonomous mode.
    • Includes steps like proposing resolution steps to decomposition of steps into child tasks (if applicable).
  • Sentiment Analysis capability enhancements
    • Ability to clone and edit sentiment skill in NASK to support extended tables functionality.
  • Record Summarization capability enhancements
    • New configuration to include AI activity like AI Specialists notes in summarization output.
• Fixed
  • Additional input fields configuration for child/copied skills are now injected into the LLM prompt for Resolution Notes Generation skills.
  • Honor "Select display" role for visibility of AI Sparkle icon for Email Reply Recommendation skills.
  • Addressed issue of non-citation of KB data for Email Reply Recommendations in spite of data being present.
  • Addressed assets getting loaded when Record summarization skills are disabled.
  • Provided an option to disable post analysis actions via skill config varset for Identify Ways to Improve Service agentic workflow.
  • Addressed multi-lingual issues with Record Field Value Prediction AI Agent for Brazilian and Portugese languages for Gemini 2.5.
• Removed
  • Stepper component for in-product experience (replaced by new interactive experience).

• Glide List Support.
• Semantic layer (Entity/Dimension) customization Support.
• Manual Segments Improvements.
• 3P Model Support (GPT 5.4 / Claude Sonnet 4.6).

Changed: OOB AI agents are made read only, users can duplicate them to modify. Specifically, appointment agent is being made read only.

• Improved keyboard shortcut labels.
• Fixed text contrast in Coral dark theme.
• Additional AI model support.

• New
  • Capture and analyze the contents of your screen directly from your browser to auto-fill form fields. To specify the area of the captured screen that you want ServiceNow AI Lens to analyze, crop the image before submitting it for analysis.
  • Set up your organization's ServiceNow instance URL once after the installation of ServiceNow AI Lens desktop app so that employees see it ready-filled when they open the ServiceNow AI Lens desktop app login screen. You can enable the auto login system property so that users can skip signing in on subsequent launches if their session is active. However, if a user's session expires or they sign out, ServiceNow AI Lens asks them to sign in again.
• Changed
  • A new screen opens when you select the Create with Lens button on a list view or Update with Lens button on a form.

Enhanced the prompts for Order Capture and Order Approval Flow.

Generate AI-powered color palettes from brand images or brand guidelines using the Now Assist panel.

Fixed (Patch Release): Service bridge name change in Agentic AI to Service Exchange.

Complaint Case AI Agents: Update fields tool allows arbitrary field modification.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Updating our dependent app versions for better packaging.

Fixed: Fixed an issue with mismatched API contracts for Intent Activity Analysis (Process Mining).

Fixed: ACLs for voice input being sent to NowLLM.

Read only fields will be now "strict" read only in general Now Assist skill tables. With this change tables cannot be updated by using client scripts, browser consoles et al.

No Features delivered via this application, it is a wrapper for the Complaint Case AI Agent collection.

Fixed (Patch Release): Service bridge name change in genai to Service Exchange.

Removing Now Assist for SOMT dependency.

• New
  • General Availability - Agent Miner graduates from Innovation Labs to GA.
  • Auto-generated evaluation datasets - Every opportunity now ships with an evaluation set for validating agent performance.
  • Opportunities list as default view - The list page is now the primary entry point for reviewing automation opportunities.
• Changed
  • Custom agent creation restricted to Prime SKU and current Now Assist packages - Foundation and Advanced tiers no longer support custom agent creation.
  • UI enhancements across the product based on customer feedback - easier navigation, clearer opportunity details, and refined settings.
• Fixed
  • Numerous UI and setup issues resolved for GA.
• Removed
  • Duplicate-and-customize existing agents - removed based on customer feedback.

• Added support for exploration of data from secure scopes.
• Enhanced query generation capabilities such as:
  • Support for global OR conditions.
  • Improved filtering on class field of CMDB tables (support for InstanceOf operator).
  • Added ability to filter by sys_created_by/sys_updated_by fields.
  • Semantic segment tuning improvements (ability to disable specific segment types).
• Administration enhancements:
  • Added module for AI Data Explorer logs (includes error and performance logs).
  • Added module for workspace configuration for AI Data Explorer.

Fixed: Inconsistencies within plugin.xml and pom files.

Fixed a bug that was causing the transcript of a conversation not to build correctly for primary request tagging.

Fixed: Set the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only. This configuration change enables seamless functionality for Advanced tier customers.

Fixed: Inconsistencies within plugin.xml and pom files.

New: Added support for Agent Advisor.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for IRM Advanced SKU subscribers.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for IRM Advanced SKU subscribers.

Fixed: Updated the prompt to resolve a tool call issue in the regulatory analysis agent.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only, supporting seamless functionality for customers on the Advanced tier.

Fixed - ACLs for voice input being sent to NowLLM.

• Changed: Updated Endpoint for Create Response Actions - The endpoint for Create Response actions now includes the version segment (v1) in the resource path, aligning with the latest Azure OpenAI API specification.
• Fixed: Data Type Format Constraint Issue for Spoke Vision Wrapper - Fixed a data type format constraint issue in the spoke vision wrapper where a field typed as object required input in a strict format - any deviation caused failures or incorrect processing.

Agentic workforce UI enhancements.

• UI size enlarged in UI16 when Polaris is switched off.
• Change summarization is displayed for a few change requests, but not for others, after the upgrade. Now Assist for ITSM to 28.7.0.
• NACM accessibility issues with header labels.

Contract entitlement data extraction skill as part of Now Assist for Software Asset (SAM).

Fixed: Minor defects.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the IRM Advanced tier.

Track which knowledge articles and catalog items support successful virtual agent deflections instead of transferring to human agents using the ITSM Virtual Agent Analytics dashboard.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the Advanced tier.

Changed: Removed install as a dependency check.

WDF Now-Assist Australia release.

Fixed: Updated the protection policy for the Suggest Potential Risks use case and the Risk Identification Agent to read-only to ensure seamless functionality for customers on the Advanced tier.

Changed: Application renamed to "Security Incident Response - Advanced".

Changed: Application renamed to "Security Incident Response - Foundation".

Changed: Application renamed to "Security Incident Response - Prime".