LendScout: Four Local AI Agents for a Mortgage Practice
LendScout is a mortgage practice in Charlotte, North Carolina. CBus AI Agents built four AI agents for the loan officer and team: a morning pipeline brief, pre-approval file scoring, follow-up drafts for quiet borrowers, and a post-close sequence. All four run as rules-based scripts on the loan officer's own computer. Learn more at lendscout-asmc.com.
Published June 2026
The situation.
A mortgage practice runs on follow-through. The loan officer needs to know where every loan sits each morning, get back to borrowers who have gone quiet, and stay in touch after closing. By hand, that means scrolling the pipeline spreadsheet and writing the same emails over and over.
Borrower files carry names, phone numbers, and financial detail. The build started from one rule: any agent that touches borrower data runs on the loan officer's own computer.
What we built.
Four agents, drafting from the practice's own email templates and voice rules. The loan officer reviews every draft and decides what to send.
Pipeline Standup
A one-screen morning brief: loans clear to close today, files that need the loan officer's attention, the under-contract watch list, files stuck in intake, and refinance opportunities worth a call.
Pre-Approval Scoring
Sorts new pre-approval files into Clean, Tight, Needs Work, or Red Flag by matching them against the team's own past file notes, then suggests the next step. The person who preps files makes the final call.
Follow-Up Drafting
Finds borrowers who have gone quiet and drafts a follow-up for each from four email templates, matched to the borrower's stage in the pipeline. The drafts wait for review.
Post-Close Sequence
Tracks an eight-touchpoint sequence for every closed loan and drafts the email touches, from the closing-day congratulations to the six-month check-in.
Privacy by design.
The agents read pipeline spreadsheets saved on the loan officer's computer and write their briefs and drafts to that same computer. Fixed rules handle the scoring, the follow-up picks, and the post-close timing.
The agents read the practice's spreadsheets without changing them, and they leave the loan system and the CRM alone. Nothing sends on its own; the loan officer picks what goes out. A handbook walks each role on the team through its daily part.
Sensitive client data?
We build AI agents that run on your own computer and leave every send to you. Book a free call and we'll show you how.
Book a free call →