A groundbreaking platform specialized for credit assessment

Our platform is built for you to make accurate credit decisions tailored after your unique customer segments within a second. The platform and the models are trained on thousands of data points and becomes smarter for each decision. It is easy to integrate to our platform through our online-API.

Updated in real-time

The difference between consumers payment behaviors today and for 10 years ago are huge. The platform is created with the knowledge that payment behaviors changes all the time and are updated in real-time to capture trends early.

Fully automated

The platform is designed to reduce manual work in favor of a more efficient intelligent system which becomes smarter for each decision. This means that you have more time to focus on your customer relationships.

Constantly optimized

The platform is built to optimize each decision after your requirements. It could be to increase conversion without increasing credit losses, or to win more application bids from agents, or understand when a customer will pay its debt.


Artificial intelligence enables smarter decisions

Our platform uses artificial intelligence (AI). This means that our models learn to find hidden rules, relationships and patterns from historical data, which are too complex for a human and traditional credit models to find.

The end product is a model, which is an expert at one pre-defined question. An example of a question could be: Will person A pay back his/her debt ? The answer of this question is based on what the model has learnt from historical data. In this case, the model predicts the answer based on what it has learnt from the historical data around how other people have paid back their debts.

A simplified example of how an AI-model can be developed has been provided. This specific model will be an expert at answering the question: Will person A pay back his/her debt? And is based on the following data:

The data that is used to generate our model consists of records with both the output value (column “repaid?) and the various input values, in reality our model has seen thousands of variables and examples. The result of an AI-model could look like the picture below, if using a decision tree.

When the model faces a new loan application will the model look at the applicants age, salary, loan amount and traverse the decision tree until it reaches one of its leaves, (“Repaid or Not repaid”)  – which will be the prediction of the model. If we exemplify with person A we see that our model had said yes – the person will pay back its debt.

We have used a decision tree in the above example, since it is easy to visualize, however it exists many different AI-models, which all have their unique benefits and ways of finding hidden patterns in your data. The models we use are much more sophisticated than the model we have used in the example above.

Evaluating the models

When we evaluate our models, we use a lot of different measurement tools to create a model, which is both accurate and robust. The most famous measurement tools we use are GINI score, ROC and TOC curves or accuracy in percentage.

Do you want to know more about our platform?

Please contact Tomas:

 tomas.sellden@evispot.ai

Tomas SelldenOur platform