How Jason Fransos Used Mortgage RAG and Automated Follow-Up for 12 Months
Mortgage questions sent Jason back through documents and previous work, while routine replies and follow-up took at least an hour every day. Patrick built mortgage RAG for guideline answers and automated text/email follow-up for client contact. Guideline lookups went from 20+ minutes to under one minute, and Jason’s contact ratio doubled.
Jason Fransos
- Built
- Mortgage retrieval-augmented generation (RAG) and automated text/email follow-up
- Sector
- Mortgage Lending
- Site
- delmarmortgage.com
Jason Fransos is a senior mortgage loan originator at Delmar Mortgage.
Questions about an active client or mortgage sent Jason back through documents and previous work. A mortgage-guideline answer could take 20+ minutes to find.
Routine replies and follow-up took at least an hour every day. That was 5+ hours across five workdays and 20+ hours across 20 workdays, before adding the time spent researching documents.
Patrick built a retrieval-augmented generation (RAG) system that finds relevant mortgage-guideline material and uses it to answer Jason’s questions. Guideline lookups went from 20+ minutes to under one minute. That made answers at least 20× faster and gave him more confidence with clients instead of digging through documents.
Docling prepared the documents for search, while pgvector made the relevant material retrievable. Patrick also used Pydantic AI, LangGraph, and FlashRAG to build the RAG workflow.
Patrick automated text and email follow-up so Jason could stay in touch with clients and nurture relationships while handling active loans. His contact ratio doubled as clients heard from him at the right time without every contact depending on manual outreach.
"My contact ratio has increased significantly because clients are hearing from me at the right time, without everything relying on manual outreach."
"Another huge advantage is how quickly I can search and find mortgage guidelines. Instead of wasting time digging through documents or second-guessing myself, I can get clear answers almost instantly, which has made me more confident and responsive when speaking with clients."
Jason Fransos
Senior Loan Originator, Delmar Mortgage
- What does mortgage RAG add beyond a regular document search?
- It retrieves relevant mortgage-guideline material and uses it to answer a question, so an originator can get a clear answer without opening and comparing documents one by one. Jason said guideline answers came almost instantly.
- Do guideline search and client follow-up have to be built together?
- No. They solve different jobs: finding answers during mortgage work and keeping client contact timely. Jason used both; another project can start with the workflow that takes the most time.
- What would a similar mortgage project need from my team?
- For RAG, the guidelines and prior-work material your team is authorized to use. For follow-up, your contact rules and text/email channels. A free audit maps both workflows before a build.
Losing Time to Document Searches and Follow-Up?
Get a free audit of both workflows.
Free 30-minute audit. No obligation, no pitch. Or take the 2-minute AI quiz.