CASE STUDY · DocuFlow AI · Business automation

From scattered inboxes to organized customer files.

A web and mobile system that collects documents from Gmail, WhatsApp and SMS, classifies them with AI and routes them to the right customer folder.

Automate your document workflow

Technical ownership: Itzhak Shaul · Architecture & hands-on development
Engagement: Selected freelance project · Client identity confidential
Core technologies: Python / Node.js · OCR + LLM · Gmail API · WhatsApp Business API · Twilio · PostgreSQL

Your customers send documents wherever it is convenient for them. Your team still needs each file in the right place. DocuFlow AI connects those two realities: multi-channel intake, AI classification and organized storage with a review path for exceptions.

A strong fit for: Service businesses and operations teams receiving customer documents across email, messaging and shared storage.

The business challenge

Documents arrive through multiple channels and in different formats. A useful automation must identify the document, extract relevant information, recognize duplicates and route the file without losing track of exceptions.

My responsibility

I designed and developed the document-processing system end to end: channel integrations, OCR and AI classification, validation, routing, and the web and mobile interfaces used to review activity.

How the system fits together

Simplified workflow illustrating the delivered capabilities.

Gmail, WhatsApp & SMS
Classify & validate
Review exceptions
Customer folders

What I built

One intake pipeline

Gmail Pub/Sub, WhatsApp Business API media webhooks and Twilio SMS / MMS bring documents from several channels into the system.

AI classification and extraction

OCR and an LLM classify invoices, contracts, identity documents and other files, and extract relevant entities.

Controlled file handling

SHA-256 deduplication and MIME validation help detect repeated files and validate file types, with a manual review queue for exceptions.

Organized destinations

Routing to Google Drive, Dropbox and SharePoint, with role-based access and an audit log through web and mobile interfaces.

ARCHITECTURE & IMPLEMENTATION

The technology, and the work behind it.

How the documented stack supports the business workflow, from the application layer to integrations and operational controls.

Multi-channel intake

Gmail API / Pub/Sub · WhatsApp Business API · Twilio SMS / MMS

Gmail events and messaging media webhooks bring incoming documents into the system. The integrations address the channels people already use, rather than requiring every sender to adopt a new intake portal.

Document understanding

OCR · LLM classification · Entity extraction

OCR and language-model processing classify invoices, contracts, identity documents and other file types, and extract relevant entities. Classification determines what kind of document arrived; extraction supplies information the downstream workflow can use.

Validation and duplicate handling

SHA-256 · MIME validation · Manual review queue

Content hashes identify repeated files, and MIME validation checks file types. A manual review queue provides an explicit place for documents that need attention instead of silently treating every file as a successful automation.

Application and data layer

Python / Node.js · PostgreSQL · React

The system combines processing and integrations with a web dashboard. PostgreSQL supports the application data layer, while role-based access and an audit log help control and inspect the workflow.

Mobile and storage destinations

React Native / Flutter · Google Drive · Dropbox · SharePoint

Web and mobile interfaces give users access to processing activity. Routing sends documents to the appropriate customer folder in the connected storage systems, bringing collection and filing into one workflow.

Follow the workflow

An illustrative walkthrough of the capabilities described in this project.

A customer document arrives through Gmail, WhatsApp or SMS / MMS and enters the intake pipeline.
The system validates the file, checks for duplicates, and uses OCR and AI to classify the document and extract relevant information.
The document follows the routing workflow to the customer folder, with manual review available for exceptions and an audit trail of processing activity.

The engineering behind the result

File automation needs more than a classification prompt. Channel handling, validation, duplicate detection and review are separate responsibilities, each with a clear place in the processing flow.

What this enables

The delivered system connects document arrival to organized storage. It provides an automated route for routine files and a visible review path for exceptions, so the workflow can be managed as a system.

BUILD A SYSTEM AROUND YOUR BUSINESS

Your workflow could be the next case study.

Picture a document workflow that starts when the file arrives, instead of when someone has time to search the inbox. I can connect your channels, classification rules and storage destinations into an automation designed around your team’s work.

Tell me which channels receive your files, where you store them and how you determine the correct customer folder. We can map the first document flow and its exception rules.

Work directly with Itzhak Shaul, founder of Methodspire and Principal Software Architect with 30+ years of hands-on experience. Start with a focused conversation about the workflow, constraints and delivery scope.

Selected freelance experience of Itzhak Shaul. Project details are summarized to protect client confidentiality.