Excel Tables — recognize, read, normalize

feld.ai recognizes tables in Excel files, understands their structure (headers, data rows, groupings), extracts the data, and converts it into a unified, normalized format.

1. Ingest Excel files — Import files via upload, email or folder watcher. feld.ai processes .xlsx, .xls and .csv with any number of sheets.

2. Detect tables — AI identifies table regions in each sheet, recognises headers, data rows and groupings — even in unusual layouts.

3. Normalise data — Column names, units and formats are mapped to a uniform target schema. Variants (e.g. "pcs", "pieces", "Stk.") are automatically unified.

4. Export structured data — Normalised data is delivered as JSON, CSV or directly via API to your target system.

Data Extraction from Files — Extract structured data from any document type.

Pricing Analytics — Compare price data from quotes and invoices.

Industry Solution: Manufacturing & Engineering — Process material certificates and service reports.

Automated Document Verification — Check extracted tables for completeness and plausibility.

Customs & Trade Documents — Read supplier declarations and certificates of origin.

How do you read an Excel table with AI?

You upload the Excel file (or send it via email, folder watcher or API). feld.ai automatically detects the table regions in each worksheet, understands headers, data rows and groupings, and outputs the data as a clean, normalized table — as JSON, CSV or directly via API. No macro and no fixed template are required.

Which AI can read and simplify Excel tables?

feld.ai specialises in the structured reading of Excel tables: the AI recognises the table structure across multiple worksheets, unifies inconsistent column names and units, and produces a simpler, normalized target table. All processing is EU-hosted on our own hardware in Austria.

Which table layouts are supported?

feld.ai recognises standard tables, nested tables, tables with merged cells and tables without clear headers. Pivot-like structures and landscape layouts are also processed.

How do I turn messy Excel data into a clean table?

feld.ai maps messy source data — mixed column names, inconsistent units, merged cells, blank rows — onto a target schema you define. The result is a clean, consistent table with uniform columns, types and units, with no manual copy-paste.

Can I prepare or export an Excel file for AI?

Yes. feld.ai prepares Excel files for downstream processing and exports the normalized tables as JSON or CSV — a clean, machine-readable format that feeds directly into downstream AI, ETL or ERP processes.

Can feld.ai process multiple sheets in one file?

Yes. All sheets in an Excel file are automatically detected and individually processed. Relationships between sheets (e.g. summary + detail sheets) are taken into account.

How does data normalisation work?

You define a target schema (columns, types, units). feld.ai automatically maps source columns to your schema — even with different labels and units. Mapping rules can be customised and saved.

What are typical use cases for bordereaux?

In the insurance industry, bordereaux arrive in hundreds of different formats. feld.ai normalises these heterogeneous spreadsheets into a uniform format — regardless of the carrier or broker.

Is there an API?

Yes. A REST API enables upload, processing and retrieval of normalised data. Webhooks notify your system when processing is complete. Batch processing is also available.