KENZEN — Intelligent receipt data extraction, expense categorization, and human validation | TREVQOO — Custom Software Solutions & AI Agents

System live in production

↗ Test the live platform I have a similar project

🔒 https://kenzen.netlify.app/kenzen.html

Expense Management

Domain

Human-in-the-Loop

Verification

In Production

Status

● ARCHITECTURE TECHNIQUE

Deployed Stack & Architecture

Engineered, built and maintained by TREVQOO

Moteur Vision Tesseract OCR & Computer Vision

Microservice IA Python FastAPI Asynchrone

Interface Vue.js & TailwindCSS

Workflow Validation Human-in-the-Loop

Conformité Journal d'audit immuable

📍 Context

Financial workflow application streamlining receipt capture, automated extraction, and expense reporting.

🎯 Need & Challenge

Manual entry of physical receipts and invoices causes accounting backlogs, reporting errors, and lost tax deductions.

💡 Deployed Solution

An automated pipeline scanning receipt images, extracting transaction dates, amounts, VAT, and vendors, followed by an intuitive human-in-the-loop review interface.

Observed Features

✓ Multi-Format Upload: mobile camera snap or digital PDF invoice ingestion
✓ Structured Information Extraction: automatic parsing of merchant identity, transaction date, gross sum, and VAT breakdown
✓ Human-in-the-Loop Verification: quick-check interface highlighting low-confidence fields for instant validation
✓ Financial Analytics Dashboard: visual breakdown of operational expenses across categories and team members
✓ Standardized Accounting Exports: export clean CSV feeds and consolidated expense summaries ready for bookkeepers

Engineering & TREVQOO Role

Document extraction pipeline design, ergonomic review interface engineering, and reliable data visualization.

Tags & Spécialités

FinTechGestion de dépensesExtraction OCRContrôle humain

I have a similar project →