Some people with no FICO scores had applications approved by Affirm’s new model.
Affirm launched a transformer-based machine learning model for real-time underwriting, approving more consumers with limited credit histories or no FICO scores. The model uses transaction-level data and credit history patterns to maintain comparable risk levels. According to Affirm, the model is live at checkout and improves explainability and decision-making speed.
How this was made
The 30-second read
Why it matters
The announcement highlights a strategic shift toward AI to capture underserved borrowers, potentially boosting loan volume.
Market read
First disclosure of a significant AI underwriting upgrade; modest trading relevance pending market reaction.
What to watch
Regulatory scrutiny of AI-driven credit decisions may delay broader adoption.
Background
Affirm announced a new AI-driven underwriting system as part of its ongoing technology upgrades.
Ticker impact
Affirm launched a transformer‑based underwriting model that approves more applications, including users with no FICO scores.
potential modest upside as investors price in higher loan origination volume.
Model improves approval rates without raising risk, but impact depends on consumer uptake and merchant integration.
Market effects
May encourage other fintech lenders to adopt advanced AI underwriting, raising competitive pressure.
U.S. consumer finance sector could see modest efficiency gains.
Limited to fintech and AI adoption trends.
Counterpoint
Model rollout could expose higher credit risk if explainability fails, leading to future loss provisions.
Key entities
- companyAffirm Holdings, Inc.
U.S. fintech lender introducing the transformer model.
- executiveLibor Michalek
President of Affirm, quoted on the new model.
