Aras InnovatorEdge AI Agents - Release Notes

The business features and benefits are as follows:

• Maintenance: Updated underlying infrastructure to maintain availability of the Impact Analysis Edge AI Agent.

The business features and benefits are as follows:

• Maintenance: Updated underlying infrastructure to maintain availability of the Smart Variants Edge AI Agent.

The business features and benefits are as follows:

• Maintenance: Updated underlying infrastructure to maintain availability of the BOM Import Edge AI Agent.

The business features and benefits are as follows:

• Maintenance: Updated underlying infrastructure to maintain availability of the Requirements Ingestion Edge AI Agent.

Aras is pleased to announce the general availability of the Impact Analysis AI Agent v1.0.0. Operating as a specialized Task Assistant Agent, this module integrates directly into your change management workspace to analyze unstructured context fields, evaluate configuration dependencies across product branches, and automate the identification of mandatory Affected Items.

Tracing downstream change propagation across deeply nested assembly hierarchies has historically forced Change Authors and Specialists into resource-intensive database lookups. The Impact Analysis AI Agent addresses this friction by reallocating human engineering hours away from manual traversal and transcription toward authoritative verification and high-value design validation. By grounding all predictive insights inside your authoritative system of record, Aras ensures that change evaluations are context-aware, reproducible, and fully auditable.

When a Change Request is submitted for analysis, the agent automatically runs five AI specialists in parallel to evaluate the change scope, examine product structure, and identify which items need to be included:

• Change Management Expert: Orchestrates the analysis and assembles the final recommendations.
• Part Model Checker: Audits part properties and structural relationships.
• Rationale Analyzer: Builds a justification for each recommended item.
• Prior Change Order Checker: Searches historical Change Orders for relevant patterns.
• ARAS Expert: Reads live data directly from your Aras Innovator instance to ground recommendations in your actual product record.

The agent guides engineers through a four-stage wizard flow designed to ensure no change is committed without deliberate human review:

1. Prepare: Load the target Change Request, review its context fields (title, description, reason for change, existing Affected Items), and optionally provide custom analysis instructions to focus the agent’s scope.
2. Analyzing: Monitor live progress as the five specialist micro-experts execute in parallel, with individual expandable progress cards and a streaming reasoning summary.
3. Summary: Review the full set of recommended Affected Items, each with a suggested action (Release, Revise, or Obsolete) and an explicit AI-generated justification. Filter and toggle individual recommendations before committing.
4. Applying: Apply selected changes back to the Change Request in a single transaction. A success confirmation and direct link back to the record in Aras Innovator are provided. Use Process Another to start a new analysis cycle without re-authenticating.

When recommendations are accepted, they are written to the native Unified Change Management (UCM) Affected Items grid. The rationale for each committed item is automatically recorded in the Aras Secure Social Visual Collaboration (SSVC) comment thread on the Change Request, creating a durable audit trail.

When the agent detects ambiguity, such as multiple historical Change Orders with divergent resolutions, it presents clickable suggestion chips directly in the Summary interface. Engineers can respond to these clarifications in a single click, guiding the agent toward higher-confidence recommendations without re-running the full analysis.

The agent is available as a software entitlement under the Aras InnovatorEdge AI Task Agents subscription model.

• Platform Dependencies: Requires that Aras Unified Change Management (UCM) be installed and active as a baseline prerequisite before configuration.
• Subscription Alignment: Requires active Aras Innovator, Aras InnovatorEdge API, and Aras InnovatorEdge AI licenses.
• Retro-Compatibility: Fully compatible with any supported version of Aras Innovator running standard UCM configurations.

To ensure that agent evaluations align with your unique data properties, custom workflows, and corporate configuration rules, the Impact Analysis AI Agent is delivered as a modular, configurable package. Deployment and customer-specific environment adaptations are fulfilled through a separately contracted Aras professional services engagement.

The business features and benefits are as follows:

• Edge AI Legal Terms: The Product Specific Licensing Terms for Aras InnovatorEdge AI are now available on the Privacy Notice and Legal Terms page.

The business features and benefits are as follows:

• Parallel import support: The agent now runs reliably across multiple browser tabs simultaneously.
• Edge AI Legal Terms: The Product Specific Licensing Terms for Aras InnovatorEdge AI are now available on the Privacy Notice and Legal Terms page.

The business features and benefits are as follows:

• Parallel Session Reliability: You can now run Requirements Ingestion imports in parallel across multiple browser tabs.
• Edge AI Legal Terms: The Product Specific Licensing Terms for Aras InnovatorEdge AI are now available on the Privacy Notice and Legal Terms page.

The business features and benefits are as follows:

• Performance Optimization: Reduction in the time required to analyze product configurations.
• Enhanced Logic Extraction: Improved co-occurrence analysis accuracy within the AI Agents Service.

___________________________________________________

This update introduces an optimized engine for the AI Agents Service, specifically tuned to handle the complex co-occurrence analysis required for variant rule extraction.

Benefits: Decreases the engineer’s wait time when invoking the “Analyze” action on multi-configuration product structures.

Example use case: An engineer managing a modular industrial machinery line can analyze dozens of valid equipment setups to automatically identify mutually exclusive sensor packages and mandatory power supply components across different product models.

The business features and benefits are as follows:

• Flexible Data Ingestion: Added support for common engineering spreadsheet layouts beyond single-column formats.
• Process Transparency: Direct capture of raw system errors during the Aras Innovator commitment phase.
• Graceful Resource Management: Improved control over background parsing jobs to ensure system resources are freed upon user cancellation.

___________________________________________________

The agent now supports “stair-step” (staircase) structural leveling in Excel files, where hierarchy is indicated by column indentation rather than a single level-index column.

Benefits: Expands the variety of supplier BOM formats that can be imported without manual file reformatting.

Example use case: A Procurement Manager can now import a legacy supplier spreadsheet where assembly levels are represented by shifted columns (staircase format) directly into Aras Innovator.

Implementation of an upgraded analysis engine to improve the throughput and accuracy of part metadata extraction from engineering spreadsheets.

Benefits: Increases the speed of the initial parsing phase for large-scale Bill of Materials.

Example use case: Ingesting a master BOM with thousands of rows will result in faster classification of part numbers, names, and quantities.

Fixed an issue where the “Cancel” button would return the user to the home page but allow the background parsing job to continue running; background tasks are now stopped immediately upon cancellation.

Resolved a display bug where raw Innovator length violations (e.g., Part names exceeding 64 characters) were hidden; clear descriptive messages now appear in the UI for remediation.

Fixed a display issue where long reference designators or part identifiers caused the Import Progress dialog to overflow its container width.

The business features and benefits are as follows:

• Reliability & Performance: Improvements to extraction speed and accuracy through an upgraded analysis engine.
• Data Integrity Safeguards: New navigation guards prevent accidental progress loss during active ingestion cycles.
• Actionable Feedback: Enhanced error processing transforms raw system technicalities into human-readable instructions.

___________________________________________________

This release introduces an enhanced backend extraction engine designed to increase the precision of requirement statement identification within complex PDF documentation.

Benefits: Improves the reliability of automated parsing, reducing the manual correction effort during the review phase.

Example use case: An engineer processing a high-density technical specification will see fewer misclassified text blocks and more accurate requirement boundaries.

To prevent data loss, the system now implements a navigation guard that monitors the browser state during active imports.

Benefits: Protects the engineer’s progress by preventing accidental page reloads or tab closures from discarding uncommitted data.

Example use case: If a user accidentally hits the browser’s “Back” button or F5 during a large document import, the application will trigger a warning to confirm the action.

This update replaces raw OData system errors with descriptive, actionable messages when a connection or system extension issue occurs.

Benefits: Allows users to troubleshoot integration issues immediately without requiring administrator intervention to read logs.

Example use case: If an Aras Innovator instance is missing the required Requirements Engineering extension, the UI will clearly state the missing dependency instead of returning an empty response.

Resolved a 404 error that occurred when users attempted to start a new ingestion cycle by clicking the “Process another” button immediately following a successful import.

Fixed a “Task Canceled” exception that caused ingestion jobs to fail when complex custom parsing instructions were applied to large-scale PDF files.

Corrected a display bug where long property names or descriptions in the import result dialog would exceed the width of the UI container, causing layout distortion.

Aras is pleased to announce the general availability of the Smart Variants Agent v1.0.0, delivered via the InnovatorEdge AI framework.

The agent ingests product Bill of Materials (BOM) data from Aras Innovator, analyzes configuration relationships across variants, and assists authors in creating a Variant Management (VM) model.

Defining and maintaining variant structures across complex product families historically requires massive manual effort.

The Smart Variants Agent reduces this friction by shifting human effort from manual transcription to high-value validation. By automating the derivation of rules from existing engineering BOMs, organizations can faster model variability structures, usage conditions, and variable components while maintaining a strict digital thread.

The agent provides a comprehensive workflow to move from raw engineering data to a verified variant model:

• Product Selection: A paginated list allows users to search for products by name, item number, or description.
• Deep BOM Expansion: The “Analyze” action loads a deep BOM expansion for the selected product and sends a normalized graph to the AI Agents Service.
• Deterministic Analysis: The agent derives variant rules through deterministic co-occurrence analysis.
• LLM Enrichment: Large Language Models are used to enrich the variant model, ensuring a robust import payload.

Before any data is committed to the system of record, engineers must perform a review of the analysis results:

• 150% BOM Tree: A merged structural view with the first two levels expanded by default.
• Rule Tables: Dedicated, searchable tables for Coupled (and) rules and Either/Or (XOR) rules.
• Dependency Matrix: An interactive matrix for Implies and Exclude rules with an expand/collapse toolbar for easier navigation.

Task Agents are available as a software entitlement under the Aras InnovatorEdge AI Task Agents subscription. Each agent is delivered as a configurable package designed to be adapted for customer-specific requirements through a separately contracted professional services engagement.

• Subscriptions: Requires active Aras Innovator, Aras InnovatorEdge API, and Aras InnovatorEdge AI subscriptions.
• Version Compatibility: any supported version of Aras Innovator (retro-compatibility)
• Applications: Aras Variant Management

• BOM Depth: The Edge App uses a fixed OData expansion depth reaching up to nine levels inclusive.
  BOMs exceeding this depth may result in incomplete analysis.
• UI Editing Constraints: The review UI is currently read-only for rules and the 150% BOM; adjustments must be made to source data in Innovator and re-analyzed to change outcomes.
• Partial Import Failure: If an error occurs in a later import phase (e.g., Phase 3), data from earlier phases may already be committed to Innovator. The UI provides error messages for remediation, such as prefix collisions.
• Single-Configuration Products: The agent requires multiple BOM configurations to produce meaningful variant rules; products with only one configuration will not infer co-occurrence relationships.
• Search Limitations: Search field inputs are capped at 100 characters to prevent fetch failures.
• LLM Transparency: The LLM-enriched model used for payload construction is internal and not reviewable in the UI; users review the deterministic rules only.

The initial Task Agent package is the first release in an expanding catalog.
Additional agents and features planned for 2026 include:

• Change Impact Analysis Agent: AI-assisted analysis of change propagation across the BOM, identifying downstream impacts on parts, documents, and processes.
• Priority Expansion: Further agents will be prioritized based on customer advisory input and pilot outcomes.

The business features and benefits are as follows:

• Requirements Import Acceleration: Reduction in time required to move requirements into Aras Innovator through parallelization.
• Enhanced Review Experience: Full-screen validation interface with persistent search and cross-tab session support.
• System Connectivity Visibility: Clear indicators of the target Aras Innovator instance currently connected.
• Integrated Compliance & Support: Direct access to Aras documentation, legal terms, and support resources.

__________________________________________________________________________________

This release introduces batching and parallelization to the Requirements Document import process. By processing 200 items per request, the agent now handles the creation, updating, and mapping of requirements to Document Items significantly faster.

Benefits: Reduces the engineering wait time for large documents.

Example use case: An engineer uploading a large technical rulebook can now complete the entire import cycle in a single session without risk of gateway timeouts.

The Review phase UI has been extended to a full-screen layout. This update includes a search field that remains visible even while scrolling through long requirement lists.

Benefits: This helps maintain search context and provides a spacious environment for detail-oriented data review.

Example use case: An engineer reviewing hundreds of extracted requirements can scroll through the entire document while keeping the search bar accessible to filter and verify specific items.

The agent now automatically cleans up the Custom parsing instructions field when a user logs out or re-logins.
Simultaneously, instructions are synchronized across multiple browser tabs during an active session, allowing users to modify logic in one tab and see it reflected in others.

Benefits: Enhances security by ensuring old parsing logic does not persist across different user sessions while supporting flexible browser-based workflows within a single session.

Example use cases: A user modifies a custom instruction in one browser tab to better segment requirement text and sees the updated instruction automatically applied when viewing the import page in a second tab.

The user interface now explicitly displays the URL of the Aras Innovator instance to which the agent is currently connected.
Information regarding the target Innovator environment is retrieved directly from the Edge Application configuration.

Benefits: Provides immediate clarity for users working in multi-environment landscapes, ensuring requirements are imported to the intended system of record.

A new Help menu (accessible via the ? icon) provides a dropdown with direct links to official Aras InnovatorEdge documentation and the subscriber support portal. Additionally, links to Aras Legal Terms and Privacy Policies are now persistently visible across all agent pages.

Benefits: Facilitates faster self-service and issue resolution while ensuring users have constant access to governed usage policies.

Fixed a critical issue where modifications made during the preview phase (such as requirement deletions or text updates) were discarded if a user switched browser tabs during the setup phase. The agent now correctly imports the updated version of requirements and the Requirement Document.

Resolved a 504 Gateway Timeout error that occurred when attempting to import large requirement datasets.

Fixed an issue where the agent incorrectly reported a “completed successfully” status with null results when an internal error occurred during job execution. The agent now correctly identifies a Failed status and displays the relevant error to the user. A Try Again button is shown redirecting back to the Upload page.

Resolved a display bug where the total requirement count incorrectly showed the initial parsed number after a successful import, failing to reflect requirements deleted by the user during the Preview phase.

Aras is pleased to introduce the BOM Import Edge AI Agent as part of the first generation of AI agents delivered through InnovatorEdge AI.

This Task Assistant Agent enriches the digital thread by transforming Bill of Materials (BOM) data contained in engineering spreadsheets into structured data managed inside Aras Innovator. The agent operates fully within the governed Aras Innovator system of record and requires human validation and review before any items are created in the database.

The agent extracts part metadata and hierarchical relationships from Excel source documents (.xlsx file type). It segments content into candidate items including Parts, Part BOMs, and associated properties, creating one Root Part (0-level) per ingestion session and recursive Part BOM relationships linking parents to children.

The BOM Import Agent supports global and custom instructions to define how spreadsheet content is interpreted:

• Natural Language Mapping: Users can describe column mappings to ensure data aligns with specific Aras properties, such as mapping “Assembly Identification” to the “Name” Part property.
• Hierarchy Detection: The agent can interpret structure via single-column level numbering (1, 1.1) or stair-step formatting.
• Uniqueness & Merging: The algorithm automatically identifies the primary (unique) key column to merge and consolidate duplicate parts within the uploaded structure.

After extraction, the agent presents a candidate structure for review:

• Hierarchical Tree Preview: View the assembled BOM in a structured grid.
• Full-Text Search: Filter and highlight text across all extracted columns to verify critical data.
• Property Editing: Modify editable Part and BOM properties (Name, Description, Qty) directly in the preview before import.

The agent identifies part images embedded in the source file and prepares them for import as Thumbnails linked to the corresponding Part items in Aras Innovator.

The key outcome of the BOM Import Agent is structured BOMs managed as governed objects. By separating content from source file formatting, teams can operate on engineering data rather than manually transcribing spreadsheet rows. This shifts human effort away from tedious transcription and toward validation and engineering intent, where domain expertise is most valuable.

The BOM Import Agent is delivered through the InnovatorEdge AI platform and operates using the Edge API, which provides the secure execution layer for AI services interacting with Aras Innovator.

• Connectivity: The Edge API manages communication between the agent runtime and the system of record to ensure all operations occur within governed platform boundaries.
• Prerequisites: The agent depends on the Aras Product Engineering (PE) application for the data model and governance framework used to manage Part and Part BOM objects.
• Supported Versions: Release 33 (or latest compatible).

• Supported Formats: Usage is currently restricted to Excel (.xlsx) files; support for PDF, CSV, or Word is currently out of scope.
• File Size: Supports a single file upload per session with a maximum size of 10 MB.
• Creation Policy: The agent creates new Parts only and does not currently support matching against or updating existing parts in the system.
• Relationship Scope: Exclusively supports Part BOM relationships; properties relying on different relationships (e.g., Cost or Weight) are not supported.
• Hierarchy Editing: Users cannot currently re-attach parts or change parent-child structures via drag-and-drop within the preview grid.
• Metadata Flags: Required properties may not be highlighted in the UI due to current Edge API metadata reporting limitations.

Aras is pleased to announce the general availability of the Requirements Ingestion Agent, the first purpose-built Task agent within the InnovatorEdge AI framework.

While conversational agents like the NLA Agent focus on answering questions, Task Agents take active engineering action to ingest, normalize, and structure product data.

The Requirements Ingestion Agent automates high-friction PLM workflows by transforming unstructured requirements trapped in PDF documents into governed items directly within Aras Innovator Requirements Engineering.

The agent enables you to ingest requirements from external PDF source documents and convert them into structured requirement objects within the system of record.

This eliminates manual transcription errors and ensures that every requirement remains traceable from the source document to the PLM record from the moment it enters the digital thread.

Manual transcription of engineering specifications has historically been a significant bottleneck in product development.
The Requirements Ingestion Agent removes this barrier by shifting human effort from tedious data entry to high-value validation and design intent.
By automating the normalization of incoming document data, organizations can accelerate their digital thread while maintaining strict requirements quality and governance.

The Requirements Ingestion Agent is available for both Full SaaS and Hybrid deployments. For Hybrid models, the Aras Innovator instance remains on-premises and connects to the Edge AI service through the Edge API layer via outbound HTTPS (port 443).

To ensure each deployment is grounded in a customer’s unique data model, workflow conventions, and business rules, the Requirements Ingestion Agent is delivered as a configurable package.
Deployment and adaptation for customer-specific environments require a separate Aras professional services engagement.

Task Agents operate within the InnovatorEdge AI ecosystem and require the following:

• Applications: Aras Requirements Engineering.
• Innovator Version: any supported Aras Innovator version
• Requirements Engineering Version: Release 33 (14.0.2).
• Subscriptions: Active Aras Innovator, InnovatorEdge API, and Aras InnovatorEdge AI subscriptions.
• Vault Access: Proper configuration of Edge API file vault settings is required to handle graphics and file uploads during the import phase.

• Supported Formats: Validation is currently restricted to PDF documents; other formats are currently out of scope.
• File Size: The agent supports a single file upload per session with a maximum size of 10 MB.
• Table Editing: Extracted tabular content is preserved and associated with requirements, but these tables are not editable during the review phase prior to import.
• Structural Dependency: The quality of extraction is highly dependent on the structural conventions (such as numbering and headings) used in the source document.

This release is the first in an expanding catalog of agents designed to optimize end-to-end PLM business processes. Future updates planned for 2026 include:

• BOM Import Agent: Normalizes incoming bill of materials data from supplier and partner formats and maps it to standard Aras Innovator structures.
• Smart Variants Agent: Analyzes existing BOMs and supporting product documentation to produce variant models and configurations within Aras Innovator.