How to build a Loan Underwriting Agent

This workflow is designed to streamline the process of generating professional loan term sheets and risk factor analyses for private lending or credit teams. It leverages AI to automate document creation and integrates directly with SharePoint for seamless document management.

1. User Input & Document Upload

• Company Name Input:
  The process begins with a simple input field where the user enters the name of the company for which a term sheet is needed.

• Financials Upload:
  Users then upload financial statements or supporting documents. The system is equipped to handle scanned documents and PDFs, using OCR (Optical Character Recognition) to extract and process the text.

2. AI-Powered Document Generation

• Term Sheet Generation:
  The workflow uses an advanced AI model to analyze both the company name and the uploaded financials. It then generates a detailed, professional term sheet. This document includes all the key sections a lender or investment committee would expect: Borrower, Loan Amount, Interest Rate, Term, Repayment Schedule, Collateral, Covenants, Fees, Conditions Precedent, and Key Risks.

• Risk Factor Analysis:
  In parallel, a second AI model reviews the same inputs to identify and summarize the top 3–5 unique risk factors for the deal. The output is concise and focused on what a credit committee would want to see highlighted.

3. Output & Document Creation

• On-Screen Results:
  Both the term sheet and the risk factor summary are displayed to the user for immediate review.

• SharePoint Integration:
  The workflow automatically creates two Word documents in your SharePoint site (“PrivateLending” on stackaintelligence.sharepoint.com):

  • One document contains the generated term sheet.

  • The other contains the risk factor summary. This ensures all documentation is securely stored, easily accessible, and ready for further editing or sharing.

4. How the Workflow Connects

• The company name and financials are both sent to the AI models.

• The outputs from the AI models are routed both to the on-screen display and to the SharePoint document creation actions.

• The SharePoint actions use a pre-configured connection to save the documents in the correct location, with clear filenames (“TermSheet” and “RiskFactors”).