Case Study
That Index Consulting×Nanonets

100+ data rules applied per invoice.
Real-time MSSQL export in 6 seconds.

That Index Consulting processed international supplier invoices via Gmail with inconsistent PO formats, no VAT validation, and no structured database output. A Nanonets agentic pipeline now ingests from Gmail, applies over 100 configurable data-action rules — including UK PO normalization, date formatting, and charity code assignment — and inserts into MSSQL in real time.

96% Time Reduction100+ Data Rules~6s Per DocumentReal-time MSSQL
About the customerThat Index Consulting

That Index Consulting is a UK-based consultancy managing supplier relationships across international markets. Their accounts team processed supplier invoices received via Gmail — in varied formats from different countries — applying UK-specific formatting rules, VAT validation, and charity code assignments manually before Nanonets automated the pipeline.

United Kingdom
HQ
Consulting
Sector
100+
Data rules applied
~6 sec
Per document

The Problem

That Index processes supplier invoices from international vendors — all arriving via Gmail in inconsistent formats. UK PO codes arrived in raw formats needing normalization. Dates weren't standardized. VAT numbers weren't validated. Charity codes had to be assigned manually per supplier. Every document required a human to apply these rules before the data could be used — and nothing was written to a structured database.

  • UK PO codes arrived inconsistently — normalization to PO-UK-XXXX format required for every document.
  • Date formats varied by country — all needing conversion to DD/MM/YYYY for UK compliance.
  • No structured output— data wasn't written to any database, making downstream reporting impossible.
Gmail
Invoice received
Manual
Download
Manual
Extract fields
Manual
Apply rules
Manual
VAT lookup
Excel
Spreadsheet
Manual process repeated per invoice
100+
Rules per invoice
~6s
Automated pipeline time
Stack
Gmail
MSSQL
Excel

The 100+ rules we needed seemed impossible to automate. Nanonets handles all of them — PO normalization, VAT validation, charity codes — and pushes the result to our database in seconds.

Director
That Index Consulting
That Index Consulting

The solution

One extraction agent. 100+ rules. Real-time MSSQL in 6 seconds.

Nanonets replaced manual rule-application with one agentic pipeline. Four agents each own a step: ingesting invoices from Gmail in real time, extracting all key fields using OCR, executing 100+ configurable data-action blocks (UK PO normalization, DD/MM/YYYY dates, GBP formatting, VAT validation, charity codes), and inserting the final record into MSSQL — completing the full pipeline in approximately 6 seconds per document.

Ingest from Gmail
Extract Document
Apply 100+ Rules
Validate & Enhance
Export to MSSQL

Pipeline overview

Gmail API
Invoice email received
Attachment auto-detected
Nanonets
Nanonets ingest
PDF pulled via Gmail API
Nanonets
Nanonets extract
>90% accuracy across templates
Nanonets
100+ post-processing rules
UK PO normalization · validation
MSSQL
MSSQL export
Real-time INSERT · 6 sec
Nanonets
Nanonets flag
Problematic template
Template exception
handling
96% time reduction · 3 min → 6 sec · Gmail + MSSQL integrated

Impact

96%
time reduction
Per document processed
100+
data rules
Applied per invoice
~6s
per document
End-to-end pipeline
9/10
satisfaction
Rated by That Index
Metric
Before
After
Improvement
Processing time per doc
Minutes — manual
~6 seconds
96% reduction
UK PO normalization
Manual per invoice
Auto-applied
Consistent format
VAT validation
Manual lookup
Auto-validated
Instant compliance
Database output
None (spreadsheets)
Real-time MSSQL
Structured at scale
Rule coverage
Human judgment
100+ rules auto-applied
Fully automated

AGENT ANATOMY

Nanonets Agent
Autonomous action-taking
After extraction
Agent decides path and acts
Real-time export
Standard invoice
Gmail API
Email ingested
Attachment detected
Pipeline triggered ✓
Metadata logged ✓
Nanonets
Fields extracted
Supplier refGGL-2024-0892
Amount£4,820.00
POPO-UK-4829
Nanonets
100+ rules applied
UK PO normalized ✓
Date DD/MM/YYYY ✓
Cross-table matched ✓
Nanonets
Validated
VATGB-829-4821 ✓
CharityCHY-4829 ✓
MSSQL
MSSQL inserted
Tablesupplier_invoices
Row847,293 ✓
Nanonets
6 sec end-to-end
Time6 seconds
Was3 minutes
Rule exception
Unrecognized PO format
Nanonets
Post-processing rule fails
⚠ Unrecognized PO format
Nanonets
Template exception flagged
StatusException queue
Nanonets
listening…
Human
Human configures new rule
New data-action block added
Nanonets
Loops back to post-processing
RuleApplied ✓
Back to pipeline
Extraction failure
Problematic template
Nanonets
Accuracy below threshold
Confidence < 90% ✗
Nanonets
listening…
Nanonets
Template flagged
QueueFlagged ✗
Nanonets
listening…
Human
Template reviewed
Model retrained on template
Nanonets
Loops back to Extract
StatusRetrained ✓
Back to Extract

See it run on your process, with your documents.

Start free. No credit card. Or talk to our team about your workflow.