Archetype (AI modelling software)
Artificially Intelligent risk modelling
- Leverage AI to create high-performing predictive risk models
- AI modelling that plays by your rules – transparent, explainable, intuitive
- Generate better performing models with a Gini coefficiency up to 10% higher
- Gain more insight from your data than ever before
- Redevelop manual and complex models more frequently
- Free up time
What is Archetype?
Archetype is our new AI-based predictive modelling suite for credit risk, fraud, and marketing. It solves the ‘black box’ problem which saw a lack of transparency with the models created, preventing the use of neural nets in credit risk.
Our patent-applied-for methodology produces AI modelling that follows sensible rules in the same way that scorecards do, but with much better results and complete transparency, which means they can be signed off through your usual governance process.
We took the development sample of an application fraud model that was originally developed using traditional logistic regression and trained a neural network on the same data using Archetype.
Archetype achieved a projected Gini uplift of up to 10% - based on our testing versus traditional regression methods using the same development and outcome data.
- Acquire more profitable customers with less risk
- Enjoy fewer losses resulting from fraud
- Benefit from improved customer targeting
An easy-to-use interface helps you:
- Upload your data
- Make decisions on what data to include
- Determine the complexity of your AI modelling
Archetype then automates previously time-consuming aspects of AI modelling, enabling your teams to focus on more complex tasks.
All you need to access Archetype is an internet connection and browser – there’s no IT installation required.
So you can quickly and effortlessly enjoy an advanced AI modelling engine using the latest techniques to execute neural net models at speed - and with a high degree of accuracy.
You may also like:
Find out more by calling our risk experts on 0333 370 6600 or email [email protected]
0333 370 6600 | [email protected]