Document Extraction Tool
Automating Book-of-Business Analysis for Transitioning Advisors

Overview

To streamline onboarding for transitioning advisors and reduce manual effort for the Business Trading Platform (BTP) team, I led the development of a Document Extraction Tool that processes screenshots and files shared by advisors, often in sensitive, early-stage transition conversations. The tool uses AWS Textract and custom logic to extract structured holdings data from unstructured files, enabling fast, secure portfolio analysis without requiring formal statements.

Challenge

Advisors considering a move to LPL often cannot share official reports from their current firm, due to internal surveillance or legal constraints. Instead, they provide a patchwork of materials including:

  • Screenshots of account dashboards
  • JPEGs or PNGs of portfolio overviews
  • Excel files with partial holdings
  • Word docs or PDFs with typed or copied summaries

The BTP team had no standardized process to extract and normalize this data. Analysts were manually interpreting each file, copying tickers and quantities, and piecing together portfolios by hand—leading to:

  • Long turnaround times
  • High potential for manual error
  • Friction in early-stage advisor conversations, where speed and trust are critical

Solution

I designed and delivered a multi-format Document Extraction Tool that automates this process end-to-end. Built on AWS Textract and post-processing logic, the tool:

  • Ingests screenshots, PDFs, JPG/PNG images, Excel sheets, and Word documents
  • Extracts structured data such as ticker symbols, quantities, account types, and asset classes
  • Cleans and standardizes the extracted data for downstream analysis
  • Outputs a CSV or spreadsheet ready for the BTP team to scope book transition needs

Key Features

  • Multi-Format File Support: Handles images (JPG, PNG), documents (PDF, DOCX), and spreadsheets (XLSX)
  • OCR + Table Detection via AWS Textract: Accurately pulls holdings data even from low-quality or cropped screenshots
  • Post-Processing + QA: Standardizes tickers, flags low-confidence fields, and prepares clean outputs
  • Secure, Lightweight Workflow: No data persistence or storage—aligned with privacy needs of transitioning advisors
  • Built for Speed: Designed to reduce time from file intake to usable output from hours to minutes

Strategic Impact

  • Empowered the BTP team to quickly scope advisor books with minimal friction or back-and-forth
  • Enabled LPL to engage earlier and more confidently with top-tier advisors in stealth mode
  • Reduced operational overhead while increasing accuracy and consistency of intake
  • Made the onboarding process feel white-glove and low-risk for advisors exploring a transition

Results

  • 75% reduction in analyst time required to process advisor-provided files
  • $250,000 costs cut in
  • Increased confidence in data used to scope book transitions
  • Contributed to pipeline growth by lowering friction in early transition conversations
  • Tool is now used regularly by BTP and being explored for expansion to other intake workflows (e.g., vendor documents, custodian files)