Realtime - underwriting
The realtime-underwriting product allows users to automate the underwriting and decisioning process for auto loans, and track lending and performance in real-time. With realtime - underwriting, users have access to a proven auto-specific risk model, trained on tens of millions of auto loan outcomes and thousands of features.
Applicant trended credit history
Deep connections between prior borrowing history and predicted future performance
Loan structure
Understanding how callback structure impacts repayment risk
Vehicle details
Identify vehicle level risk factors
Macro economic factors
Include market conditions into risk prediction
How customers use realtime - underwriting
Optimize your underwriting using our AI trained on millions of auto loan outcomes.
Full automation
Realtime-underwriting is used for nearly all auto lending efforts, with minimal human intervention on callbacks. Customers typically work up to this level in a phased approach.
Partial automation for underwriting
Realtime-underwriting is used to identify a target set of applications for automation, usually a segment outside the current buy-box and one where our models are shown to win high performing accounts.
-1.jpg)
Increase win rate
Realtime-underwriting is used to optimize callbacks to increase capture rate while mitigating risk.
-2.jpg)
Increase yield
Realtime-underwriting is used as a powerful layer of risk mitigation, where the AI underwriting model is set to a strict maximum threshold and the callback optimization model is set to yield maximization setting.
-3.jpg)
Fast adjustments to underwriting criteria
Unlike traditional auto lending, adjustments to program guidelines or realtime-underwriting model thresholds are same day processes, allowing customers to quickly react to changing lending conditions.
-4.jpg)
Challenger model
Assess benefits of using true AI underwriting vs. manual performance. Identify new segments of the credit spectrum where opportunity exists to win good credit with structured callbacks.
-5.jpg)
How do we integrate
After set up is established, Karus recommends a trial period of test applications to ensure all configurations are set correctly, and to optimize callbacks for customer specific markets.
Through an existing LOS partnership
Applications are routed to real-time through an existing integration. realtime-underwriting executes risk and callback models, and populates decisions for user review in the LOS interface. Credit pulls are typically handled through the LOS and the information is passed in the application to realtime-underwriting.
Custom API
Realtime-underwriting and customer technical teams set up a direct API connection, where applications are posted to realtime-underwriting endpoint with a callback decision as the response. Credit pulls can be handled by realtime-uderwriting or passed in the application by the customer to the API.