You’re offline. This is a read only version of the page.
10
Hi Everyone,
I’m currently working on an invoice automation and reconciliation solution using Power Automate + AI Builder/OCR, where invoices are processed and converted into structured JSON for downstream SAP reconciliation.
We are facing major challenges because invoices come from more than 1000+ vendors, and every vendor has a different invoice layout, table structure, font style, alignment, tax section, and page format.
Our current production accuracy is around 60%, which is creating significant manual validation effort.
Some common issues we are facing:
Incorrect extraction of invoice number, GSTIN, dates, and totals
Line item tables breaking across pages
Multi-line descriptions shifting columns
Different tax structures (CGST/SGST/IGST/TDS)
Low-quality scans and image-based PDFs
OCR reading incorrect numeric values
Header/footer duplication on multi-page invoices
One critical issue:
In some invoices, the OCR/model reads quantities incorrectly. For example, it gets confused between values like 5 and 6, along with several similar numeric recognition issues.
This causes reconciliation failures and incorrect invoice posting.
Currently, we are mainly using:
AI Builder “Get text from document”
Power Automate parsing
JSON mapping logic
However, since this is primarily OCR-based extraction and not vendor-trained document understanding, accuracy varies heavily depending on invoice quality and layout.
I would appreciate guidance or suggestions on:
Best approach for handling 1000+ vendor invoice formats
Improving OCR accuracy in production
Preprocessing techniques before AI Builder
Better approaches for table extraction
Whether Azure Document Intelligence/Form Recognizer performs better than standard AI Builder OCR
Hybrid approaches using OCR + LLM + rule engine
Confidence-score-based validation strategies
Recommended architecture for scalable invoice-to-JSON extraction
Also, if anyone has articles, documentation, benchmarks, or production learnings related to AI Builder extraction accuracy and large-scale invoice processing, that would be extremely helpful.
Would really appreciate recommendations and insights from teams handling enterprise-scale invoice processing systems.
Thanks in advance.
Your current setup is hitting limits because AI Builder OCR isn’t designed for high-volume, multi-vendor invoice formats, which is why accuracy is stuck around 60%.
The most effective improvement is to switch to Azure Document Intelligence, which handles invoices much better, especially tables, multi-page documents, and gives confidence scores for each field. You don’t need to train everything,just focus on your top vendors first.
To improve accuracy further, add a preprocessing step (better image quality, 300 DPI, contrast fixes) to reduce numeric errors like 5 vs 6 confusion.
Then build a hybrid flow: use Document Intelligence for extraction, an LLM to clean up messy line items, and rule-based logic for tax and SAP mapping. Add a confidence threshold, so low-confidence extractions go to manual review instead of SAP.
Also, route invoices based on vendor so you’re not treating all formats the same.
Overall, this approach typically pushes accuracy to 85–90% straight-through processing, with the rest handled by human validation.
Share
Report
All responses (
Answers (
