Ramkumar Venkataraman
CTO & Co-Founder
62 articles
Mortgage Fraud Detection at Origination: What an AI Agent Can Flag, What It Cannot Decide, and Where the SAR Obligation Starts
Most origination fraud is not a forged document, it is a set of facts that are each plausible and collectively wrong: an owner-occupancy claim that does not fit the file, an employer that only exists on paper, a gift that is really a loan. Where an AI agent reads the whole file for the pattern instead of each document in isolation, what FinCEN's mortgage AML rule and the Red Flags Rule actually require, and the line between a fraud flag and a fraud determination.
Verification of Employment With an AI Agent: The Written VOE, the Day-of-Closing Verbal, and Why The Work Number Is a Credit Report
VOE reads like a phone call and is actually three separate controls with three separate failure modes. The written verification, the verbal within ten business days of the note, and the reverification when something changes. Where an AI agent runs each one, what Fannie Mae B3-3.1-07 actually requires, and why treating a database VOE as free of FCRA is the mistake that shows up in a dispute.
The Large Deposit Question: Verifying Assets With an AI Agent Under Fannie Mae B3-4.2-02, and Where Asset Sourcing Meets the Bank Secrecy Act
Asset verification looks like adding up account balances and is actually a sourcing investigation. The 50 percent large-deposit rule, funds that have to be the borrower's own, gift documentation, and the point where an unexplained deposit stops being an underwriting condition and becomes a source-of-funds question a regulated lender cannot ignore. Where an AI agent reads bank statements, what it flags, and the line between an eligibility problem and an AML problem.
Reading the AUS Findings: Where an AI Underwriting Agent Adds Value, Where SR 11-7 Draws the Line, and Why the Model That Touches Credit Is Governed Differently
The automated underwriting system already returned Approve/Eligible. The work that follows, clearing the findings, reconciling the conditions against the file, and deciding what the AUS could not see, is where cycle time and defects live. Where an AI agent operates on DU and LPA findings, why an agent that influences a credit decision falls under model risk management, and the governance we run so the agent is a documented, tested model and not a black box in the underwriting path.
AI in Appraisal Review: Reading the Report, Triggering the ROV, and Never Touching the Value
Collateral review is where AI can help most and overstep worst. How to build an appraisal-review agent that flags quality and bias risk, routes reconsideration-of-value requests under the 2024 interagency guidance, and leaves the opinion of value to a licensed human.
Mortgage QC at 100% Coverage: What Changes When AI Reviews Every Loan Instead of a 10% Sample
GSE quality control rules were written around sampling because full review was impossible by hand. AI removes that constraint. What a full-population pre-funding and post-close QC program looks like, and where the Fannie and Freddie requirements still bind.
AI Income Calculation for Self-Employed Borrowers: The Part of Underwriting Where the Math Has to Be Auditable
Self-employed income is the hardest number in a loan file and the easiest one to get wrong. How to build an AI income engine that matches Fannie Mae Form 1084, holds up under ATR/QM, and carries an audit trail an underwriter and a model validator both trust.
Calculating Self-Employed Income With an AI Agent: The 1084 Cash-Flow Analysis, the 4506-C Transcript, and the Reasonableness Call the Underwriter Owns
Self-employed income is where mortgage underwriting is slowest, most inconsistent, and most exposed to fair-lending risk, because two underwriters can read the same tax returns and reach different qualifying income. An AI agent can run the Form 1084 cash-flow analysis the same way every time and document every add-back to its line on the return. What it cannot do is make the reasonableness determination Fannie assigns to the underwriter, and building the agent so it stops at that line is the whole design problem.
AI on the Collateral Desk: The AVM Quality-Control Rule Now in Effect, the Reconsideration-of-Value Process, and the Appraisal-Independence Line the Agent Cannot Touch
Two things changed the collateral desk in the last two years: the interagency AVM quality-control rule that took effect October 1, 2025, and the interagency reconsideration-of-value guidance finalized in July 2024. Both put new obligations on how lenders use automated valuations and how they let borrowers challenge an appraisal. An AI agent can run the collateral review and the ROV intake at volume, but appraisal independence draws a hard line around what the agent is allowed to do to a valuation. Where the agent sits, and where it has to stop.
The AI Agent Behind the Mortgage Point-of-Sale: URLA Intake, the TRID Application Trigger, and Preventing the Conditions Before Underwriting Ever Sees Them
Most mortgage point-of-sale tools collect a 1003 and stop. The work that decides cycle time happens one layer down: reading what the borrower entered, catching the missing document while the borrower is still in the session, and knowing the exact moment intake becomes a TRID application with a three-day disclosure clock attached. Where an AI agent sits in the POS, what it is allowed to decide, and the compliance lines it cannot cross at the front door.
Pre-Funding QC With an AI Agent: The Fannie D1-2 Review, the Defects You Catch Before the Wire, and the Reverifications the Rule Still Wants a Human to Judge
Post-closing QC tells you how many defective loans you already sold. Pre-funding QC is the only review that changes the outcome, because it happens before the money moves. An AI agent can run the full-file pre-funding review at 100 percent of the pipeline instead of a sample, catch the income-calculation error and the data-integrity break before closing, and drive the reverifications the rule requires. What the agent computes, what stays human, and how the defect taxonomy has to be built so the review is defensible.
NACHA WEB Debit Account Validation and the Credit-Push Fraud Vector: The Rule the ODFI Signs, the RDFI Reads, and the AI Verification Layer That Actually Reduces Return Rates
The NACHA Operating Rules for WEB debit entries require a commercially reasonable fraudulent-transaction detection system that validates the receiving account is a legitimate open account before the first ACH debit. Credit-push fraud, which bypasses the WEB rule entirely by tricking the sender into originating a legitimate ACH credit, is the fastest-growing ACH fraud vector. The rule mechanics, the account-validation architecture we run, and the counter-controls on the credit-push side of the ledger.
You Ain't Seen Nothin' Yet
- Any loan type, any agency guideline or custom investor overlays.
- Every finding cited to the guideline or document it came from