Ingestion Agent

This agent extracts text from uploaded expense documents, such as PDFs or scanned receipts, using Google Document AI. It applies OCR to the raw content and makes the extracted text available for validation and analysis.

Once a user uploads a document, the agent reads it automatically, recognizes the input, extracts structured text, and stores it in shared context for downstream agents. It is part of Expenselytics, which addresses manual expense validation that is error-prone, slow, and leads to policy violations as finance teams parse receipts and go back and forth with employees. Automating text extraction reduces manual processing time and errors, enabling faster, more accurate validation at scale, along with cost savings and better compliance. The agent runs on Google Document AI, Python, and cloud storage.