Credit Risk

Filter by:
Loan Origination: The Creditworthy Applicants Your Model Cannot See

Retail Banking  ·  7 min read

Loan Origination: The Creditworthy Applicants Your Model Cannot See

Better underwriting on a $1B loan portfolio is worth $2–5M in avoided losses annually. That is the cost of the decision quality gap on the population your model can already see. The larger and less visible cost is the creditworthy population your model cannot see at all — and the digital-first lenders that are approving them instead.

12 May 2026 Loan Origination Credit Risk
Mortgage Delinquency: The Intervention That Works Is the One That Happens Before the Missed Payment

Retail Banking  ·  7 min read

Mortgage Delinquency: The Intervention That Works Is the One That Happens Before the Missed Payment

Early delinquency signals in mortgage portfolios are detectable in behavioral and transaction data weeks before a payment is missed. The intervention options available at that point are substantially broader, cheaper, and more likely to succeed than those available after the first missed payment. Most banks are not intervening at that point because they are not identifying the risk until it surfaces.

12 May 2026 Mortgage Servicing Delinquency Prediction
Mortgage Underwriting: When Two Underwriters Reach Different Conclusions on the Same File

Retail Banking  ·  6 min read

Mortgage Underwriting: When Two Underwriters Reach Different Conclusions on the Same File

Manual mortgage underwriting introduces variability that has a cost in both directions. A creditworthy borrower declined by one underwriter would have been approved by another. A marginal borrower approved on a subjective judgment would have been declined on a consistent model. AI-assisted underwriting does not replace underwriter judgment. It makes it consistent.

12 May 2026 Mortgage Underwriting Credit Risk