I’m looking for a simple proof-of-concept that lets me drag-and-drop a customer’s full loan file—PDFs and JPGs together—and have an AI engine instantly read everything, then output four key items: • Loan eligibility based on the criteria I’ll supply the criteria • A checklist of any missing documents. • A list of incomplete form fields. • Likely follow-up queries an underwriter might raise. A single shareable report—ideally a PDF that I can forward on WhatsApp—will be perfect for now. Keep the scope lean: basic OCR for mixed files, quick rule-based or lightweight ML logic for the eligibility calculation, and a clean, minimal interface. A small web app, Google Colab notebook, or lightweight desktop utility is fine as long as it works end-to-end and I can test with sample files. Please include: • A straightforward way to update or add eligibility rules later. • Clear setup instructions. I don’t need flashy visuals yet—functionality first.