The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Quantity and unit price can swap places without changing an invoice total.
At Procys, my document AI work connects OCR and LLM layout reasoning with deterministic validation, field confidence and exception review.
Here's a test case I'd include before connecting a new supplier layout to accounting software. A synthetic invoice shows 2 units at £50 each. The extractor swaps the fields to 50 units at £2. Both produce £100, so a line-total check alone misses the error.
I'd check quantity and unit price against the source columns, then route an ambiguous row to a reviewer with its document context. The acceptance check should cover what the next system will do with each field.
Public case: https://kapilchauhan.netlify.app/case/procys.html
kapilchauhan.netlify.app
Procys: Document intelligence | Kapil Chauhan
Production LLM agents for document processing and identity verification
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
CRAG — Self-Correcting Research Assistant
Overview
A research-assistant project exploring how AI applications can retrieve information and improve the quality of their responses through a self-correction workflow.
Project focus
Python-based AI application development.
Retrieval and research workflow design.
Improving answer reliability through evaluation and correction.
Technologies: Python, AI, retrieval-augmented generation.
Self correction is where RAG gets useful. How do you judge a retrieved chunk as bad enough to trigger a new search?
Title: Data Formatting & Cleaning in Excel
Description: I turn messy spreadsheet exports into clean, consistent, analysis-ready data. I remove duplicates and extra spaces, standardize names, dates and currency, and flag missing values for review instead of guessing.
Tags: Data Cleaning, Excel, Data Formatting, Data Entry, Spreadsheets, Virtual Assistant
Your competitors can change their prices overnight. Your data collection system shouldn't miss it. 📊
Imagine checking 50 product pages every morning just to find out which prices have changed.
Now imagine getting that information automatically.
With Python web scraping and automation, businesses can build systems that:
→ Monitor product prices across multiple websites
→ Track stock availability
→ Compare competitor pricing
→ Record changes over time
→ Send alerts when something important changes
→ Export updated data to Excel or CSV
Instead of repeatedly checking websites, you get organized information that helps you make better business decisions.
This can be especially useful for e-commerce businesses, online retailers, and teams that regularly research competitors.
At ScrapePro Solutions, I build custom Python scraping and automation solutions tailored to specific business needs.
A quick question for e-commerce business owners:
If a competitor changed their product price today, how would you find out — manual checking or automatic alerts?
💬 Share your answer in the comments.
Two things bit me while building a price tracker. A page can come back "successful" but empty (in my case, compressed content the client couldn't decode), and saving that blindly overwrites good data. So I refuse any page with no title and no price, and only record a change when...