AI SharePoint Workflow Agent

An intelligent document automation agent built on Azure OpenAI that classifies, summarizes, and routes government contract documents through SharePoint pipelines without human intervention.

Project 4 cover

Overview

Developed for an AI software consulting engagement, this agent monitors a SharePoint document library for incoming RFPs and contract files, uses Azure OpenAI with retrieval-augmented generation to produce structured summaries and extract key deadlines, then routes each document to the correct team channel via Power Automate — with status updates written back to the SharePoint list automatically.

Technologies

Python · Azure OpenAI (GPT-4o) · Azure Functions · SharePoint REST API · Microsoft Power Automate · Azure Blob Storage · LangChain (RAG pipeline) · custom prompt engineering

Outcome

Eliminated roughly 60% of manual document-handling work, allowing the team to process twice as many RFP cycles per quarter at the same headcount. Average time from document receipt to routed summary dropped from several hours to under four minutes, and classification accuracy across 14 document categories held above 91%.

Detail view A
Detail view B

Left: end-to-end workflow diagram showing the SharePoint document library trigger, Azure Function handler, RAG-based classification and summarization step, and Power Automate routing logic that delivers outputs to the correct team channel. Right: the Azure Functions monitoring dashboard displaying invocation logs, processing latency per document, and error-rate tracking across a two-week production window.