Atidot developed the first ever programming language for insurance. Our DSL (domain specific language) is using insurance terms and turns them into code.
Querying "premium paying policies" could be a month of research or a phrase that the compiler analyses for you. The scenario that you build with LIA is human readable, can be complied into flow charts and can be used to explain the model to your team, to your client and to your regulator.
There are no black boxes!
This code allows us to benchmark behavior between products, companies and channels to model the most accurate interaction between the client and the policy.
Micro segmentation and lookalikes
Our models generate micro segments of people that interact with the insurance in the same way. The concept of lookalikes allows you to understand your customers and their risk profiles much better.
Everybody knows intuitively that not all your classmates are like you although you are the same age and gender. Some have more/less kids, some have more/less money, some like to travel and some take care of their parents.
If you can segment your customers by 500-1000 parameters you get the most accurate prediction of customer behavior.
Risk, profit, EV, LTV, premium, face amount
Data scientist will talk to you in their language - accuracy, precision, recall, true positive, beta errors, feature importance.
This is all very important but this is not what you need to run your business. Atidot is building a platform for the insurance industry. We are committed to translate this data science terminology into the terminology you understand.
Turning prediction to lapse assumptions, premium payment patterns, under insurance amounts, product suggestions and customers' simulation by EV or face amounts.
Our dashboards are readable to the business user to help them run their business using top of the art models and technology.
Do you want to see it in action?