Master audits before you originate

If you can’t explain your ATR/QM rationale and document TRID timing, you’ll fail internal audit/QC long before any examiner arrives. In last month’s 10-file QC pull, we flagged HMDA LAR inconsistencies and missing LOEs that a 15-minute pre-funding checklist would have caught. What training prepares new LOs for pre-funding and post-close audits — mock audits, LOS workflow controls, or something similar — and verify anything you adopt against your company policies and the regulations.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‍​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍​⁠‌​‍‍‌‍‌⁠‌⁠​​​⁠‍‌‌​‍​‌‌‍‌‌‌​‌‌⁠​‍‌⁠​​​⁠​​‌⁠‍​‌​​⁠‌‍​⁠‌‌‌⁠‌‍⁠‍​‍​‍‌⁠⁠‌

I learned the hard way that a 10-minute ‘pre-fund sanity check’ saves QC heartburn: I built an LOS task that forces ATR/QM rationale notes, TRID timestamp screenshots, and a quick HMDA LAR compare before docs. Mock audits are great, but pairing a new LO with a closer for a week shows where LOEs and timing get missed in the workflow. ‘If it’s not in the file, it didn’t happen’ has kept me clean in audits.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‍​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‍⁠‌⁠​‌‌​⁠‌‌‍⁠​‌‌‌​​⁠‌​​⁠​‌‌⁠​‌‌‍‌⁠‌‌‍​‌‌‌‌‌⁠‌⁠‌‌​‍‌‍‌‌‌​‍​‌​⁠‌​‍​‍‌⁠⁠‌

Building on @c_gray89, the best prep I got was a rotating 10-file mock audit where you have 60 seconds to defend ATR/QM per loan. I now keep a one-page “ATR/QM rationale + LOE index” in the LOS and run a pre-funding HMDA LAR export to catch ethnicity/rate spread blanks — a “15-minute pre-funding checklist” that works. Caveat: if your LOS won’t export LAR mid-stream, mirror the LAR fields in a simple sheet and make it a required task before CTC.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‍​⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍​⁠‌​‌‌‌​⁠⁠‌​⁠‍‌‍​‌‌‌⁠⁠‌‌‍​‌⁠​⁠‌⁠‍​‌​‍​​⁠​​‌​‍​‌‌​‍‌‍⁠‍‌‌‍​​⁠‌‌​‍​‍‌⁠⁠‌

, this drives me nuts — I stole @c_gray89’s 60-second defense idea and turned it into one LOS gate: an “Audit Ready” task that blocks docs until I drop TRID timing screenshots, the lock history PDF, and a canned LOE for deposits. That alone killed most of the “missing LOE” and date mismatch hits we saw in our last 10-file pull. If your LOS can’t hard-stop, a shared folder with strict file names is a decent stopgap.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​​​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‍​‌‍‍‍​⁠‍​​⁠​‍​⁠‍​​⁠‌⁠‌​⁠‌‌‍‌‍‌⁠​⁠‌‌‍​‌​‍⁠‌‍‌‍‌‌‌​‌⁠​⁠‌‍​⁠‌‌⁠⁠​‍​‍‌⁠⁠‌

I build a single “Audit Packet” PDF pre-funding: TRID date screenshots + fee change notes; quarterly peer mock audits.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​​​⁠‍​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠‌​‌‍⁠⁠‌‌‌⁠​⁠​‌​⁠‌⁠​⁠​‌‌⁠‍​​⁠‌​‌⁠‌‍​⁠​‌‌​⁠​‌​‍‌‌​​‍‌‍​‌‌‌‌‍​⁠‍​​‍​‍‌⁠⁠‌

We had one slip a day last week because someone treated signing day as Day 1; now we text borrowers “Sign Mon = rescind Tue–Thu, fund Fri” and timestamp CD receipt in Encompass so everyone’s counting from the same point. > exact starting point of the rescission period, with members sharing their interpretations and experiences. This topic is crucial for ensuring compliance and 100% — this drives me nuts, so we anchor to the day after consummation per our investor guide and post the funding date in the file header. Small caveat: some investors handle local holidays or time zones differently, so double-check their calendar before you promise a.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​‍​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​⁠​‌‍‌​‌‍‍​‌​‌‌‌​‌‍‌⁠‌​​⁠​‌‌‍‌‍‌‍‍⁠‌⁠‍‍‌‌‍‍‌​‍⁠‌​⁠​‌‍​‌‌​‌​‌⁠​‌​‍​‍‌⁠⁠‌

I’ve cut surprises by dropping a single “rescission timeline” note in the file with the eSign audit (e.g., Consummation 11/24 4:20pm CT; NORTC at table; days 11/25, 11/29, 11/30; target fund 12/2) and texting that end date to the borrowers. If a non-borrowing owner was missed, we reset the clock and ping @funding.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​‍​⁠‌‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​‍⁠‌‌‍⁠​​⁠​‌​⁠‌‍‌‌​⁠‌⁠‌​‌‍‌​‌⁠‌⁠‌​​⁠‌⁠​‌‌​‌⁠‌⁠‍‍‌​‍‌‌⁠‍‌‌‍‍‍‌‌‍‍​‍​‍‌⁠⁠‌

I pin a note in the borrower portal with a bold ‘Earliest fund: Tue 3/12 9:00am PT’ and add one line on why the window lands there — saves us from 2 a.m. time‑zone gymnastics. I also trigger an LOS task 30 minutes before rescission ends so docs/conditions are staged, but double‑check escrow and warehouse bank cutoffs since they don’t all run on the same clock.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​‍​⁠‍​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌‍​​⁠‍​‌‌‌⁠​⁠‌‍‌​​‌‌‍‍‌‌‍​‌‌​‌⁠​⁠‍‌‌⁠​⁠​⁠‌‌‌‍​‌​⁠‌​​‍⁠‌‌‌‌⁠​⁠‌‌​‍​‍‌⁠⁠‌

Quick tip from training days: we run mock audits where new LOs write a two‑sentence ATR/QM “story” and attach the income calc worksheet, then a peer uses a 10‑item checklist to spot missing LOEs and HMDA mismatches. It works only if that checklist maps to your actual QC defect codes; otherwise it’s just trivia. Think math class: “show your work.”.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍⁠⁠‌‍​‌‌‍⁠‍‌‍⁠⁠‌‍‌‍‌‍‌‍‌‍‍‌‌‍​⁠‌‍‌‌‌⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​‍​⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠​⁠‌‌​​‌‌‍‌​⁠‌‌‌‌​‍‌‍⁠⁠‌‍‍⁠‌‍​‍‌​​‍‌‍‌‍‌‍‍‌‌‌​‌‌‌‌⁠‌‌​⁠‌‍⁠⁠‌‍‌‍​‍​‍‌⁠⁠‌