Bill of materials extraction
BOMs are the heaviest documents in procurement: hundreds of lines, part numbers, hierarchy. ProcureMind extracts every line of a bill of materials into structured, reviewable data. No retyping Rev C by hand.
Start freeFields built for BOMs, not borrowed from invoices
A bill of materials isn't a quote with more rows. It has its own anatomy. ProcureMind's extraction schema carries BOM-specific fields other tools flatten away:
| BOM field | What ProcureMind captures |
|---|---|
| Manufacturer + MPN | manufacturer and manufacturer_pn capture exactly who makes each part and its manufacturer part number: the two fields that matter when sourcing. |
| Reference designators | reference_designator maps each line to the schematic: "R1, R2, C3". Essential for PCB and assembly BOMs. |
| Hierarchy levels | bom_level preserves the structure: top-level assemblies (1), sub-assemblies (1.1), and the parts beneath them. |
| Revision | bom_version records which revision of the BOM this extraction came from: "Rev B", "v1.2". |
| Quantities and packing | quantity, unit, MOQ and standard pack quantity per line, so ordering maths survives the trip. |
| Lead times and notes | lead_time per line, plus a notes field for line-level remarks. |
Built for the way BOMs actually arrive
As Excel files, mostly. Engineering exports a BOM from the CAD tool, purchasing receives it as .xlsx. Drag it into the ProcureMind dashboard, or, if it arrived as a Gmail attachment, extract it without leaving the inbox. CSV and PDF work too, for suppliers who insist on them.
Every line, individually. BOM extraction doesn't summarise. Each row becomes its own structured line item: description, SKU, manufacturer, MPN, designators, quantity, packing, lead time, so a 400-line BOM comes back as 400 reviewable items, not a blob of text.
Reviewed, then yours. The extraction opens in an editable review panel with a confidence score. Spot-check the lines the AI flagged, correct in place, and save. The full structured BOM lives in your extraction history, tagged document_type: "bom", ready to filter, export or pull over the API.
Frequently asked questions
Can ProcureMind extract a BOM from an Excel file?
Yes. Excel is the natural home of most BOMs, and it's a first-class input. Upload an .xlsx, .xls or .csv (up to 10 MB) in the dashboard, or extract an Excel attachment straight from Gmail. The response even reports how many sheets and rows were read.
Does it handle multi-level (indented) BOMs?
Yes. The bom_level field preserves hierarchy depth: a line at level 1.1 sits under its parent assembly at level 1, and that structure survives extraction into your dashboard record.
What about manufacturer part numbers?
They're a dedicated field. manufacturer_pn captures the MPN separately from any internal SKU, and manufacturer captures the maker: Texas Instruments, Murata, whoever the line specifies. No more MPNs lost in a description column.
Does a BOM extraction include prices?
BOMs aren't pricing documents, so the financial fields come back blank by design. If a supplier sends a priced BOM, the unit price and line total columns are still captured per line. The schema carries them when they exist.
How do I check the extraction before trusting it?
Every line item lands in the review panel first, with a confidence score on the extraction. Scroll the rows, fix anything the AI misread (a misaligned column, an ambiguous unit), and only then save. With BOMs running to hundreds of lines, that review step is the difference between useful data and a cleanup job.
More extraction workflows
Supplier quote extraction
23 fields per quote: vendor, terms, line items.
Purchase order extraction
PO numbers, delivery addresses and terms.
RFI extraction
Requests for information: questions and deadlines.
Gmail integration
Extract BOM attachments without leaving your inbox.
Extraction guide
The complete how-to: from copy-paste to automatic.
API reference
The LineItem schema, field by field.