BOM Import Agent - User Guide

The BOM Import Agent is a Task Assistant Agent within the InnovatorEdge AI framework.
It enriches the digital thread by transforming Bill of Materials (BOM) data contained in engineering spreadsheets into structured, governed Items within Aras Product Engineering (PE).

The agent operates directly inside the system of record to normalize incoming supplier data. Its purpose is to:

• Extract part metadata and hierarchical relationships from Excel files.
• Convert unstructured or semi-structured spreadsheet rows into governed Part and Part BOM Items.
• Reduce manual transcription effort for large, multi-level structures.
• Establish traceability by linking imported items to the authoritative digital thread.

The agent is user-invoked and human-validated. No data is committed to Aras Innovator until the user confirms the import summary.

The quality of extraction depends on the structural conventions of the source Excel file. The agent analyzes:

• Hierarchy Indicators: Single-column leveling (1, 1.1) or stair-step indentation across columns.
• Property Labels: Header rows defining Part Numbers, Names, and Quantities.
• Metadata Consistency: Use of consistent units of measure and procurement types.

Constraints:

• Format: Supports a single .xlsx file.
• Size: Maximum file size is 10 MB.
• Structure: The file must define exactly one Root (Level-0) Part.

Agent Instructions configure how the AI interprets specific spreadsheet columns. These apply globally across the document during the ingestion phase.
Users can provide natural language rules to align source data with the Aras Data Model:

• Mapping Example: “Map ‘Assembly Identification’ column values to ‘Name’ Part property;"
• Mapping Example: “Map ‘Actual’ column values both to Part and BOM ‘Status’ properties.”

During extraction, the agent identifies and segments the following content:

• Part Metadata: Unique identifiers (item_number), names, and descriptions.
• BOM Relationships: Parent-child links (source_id to related_id), quantities, and find numbers.
• Graphics: Embedded images are identified and prepared for import as Part Thumbnails.

After extraction, the agent presents a structured Tree Preview for review. This phase allows engineers to apply domain expertise before system commitment:

• Hierarchical Search: Filter and highlight text across thousands of rows to verify critical items.
• Inline Editing: Modify editable properties (Description, Quantity) directly in the grid.
• Visual Review: Verify extracted thumbnails against part descriptions.

Once validation is complete, the agent executes an atomic import sequence:

1. Part Creation: Creates all new Part items identified in the structure.
2. Relationship Binding: Establishes the Part BOM links using the verified hierarchy.
3. Thumbnail Upload: Attaches extracted graphics to Part records.

Clicking the Root Part link opens the standard Aras Innovator interface. Here, you can see the Part Form populated with the name, description, and thumbnail extracted by the agent.

Navigate to the BOM Structure tab on the Part record to view the multi-level hierarchy as it exists in the system of record. Every link, quantity, and sequence created by the agent is now part of the authoritative product record.

• Review AI Confidence: Pay close attention to items where data was semi-structured via custom instructions.
• Human-in-the-Loop: All AI-generated outputs must be reviewed by a qualified engineer before system commitment.
• Resolution Mode: Ensure you are logged in correctly; your import will respect your native Aras Innovator permissions.