Scale | InsurTech
***
Hi everyone
My name is Filipe Charters, and I’m the CEO of Data XL
We love the insurance sector; we believe that insurance companies have much to give to the world
Our business model is to provide tools - essentially insurance pricing tools – that make insures rock.
Today we bring to you our two flagship products.
. SUPPLIER OVERCHARGE DETECTION – a special case of fraud.
. Price optimization with an API
An optimal price is a price that optimizes the margin, the revenue OR minimizes the churn.
Insures companies define price in the following way:
(SLIDE)
. They sum all costs, and they say they a price. When you sum all costs, you have the total cost.
. insurers put the price near the cost, we want to put the price near the value.
This is why an IPHONE costs twice as much as a Samsung, even though the raw material is the same.
(SLIDE)
So, we build this platform, where insurance companies can put all their renewals information and receive the optimal price back.
The same input with same extra columns:
- The price that max the revenue
- The price that max the margin
- The price that minimizes the churn
This platform is working now. It’s secure – it’s cloud base in a world class provider (digital ocean), but it can be stuck in your servers.
We do not define the strategy!! So, I cannot talk about the financials, but we can imagine what would mean for a price increase of 1%, maintain the same renewal rate.
As to accuracy we can say that our margin of error was more or less 2 %.
Just 2%.
(SLIDE)
Why are we different?
To make this happen we need:
. the risk numbers (which insures already have), the price of the competition and the will 2 pay model.
We developed some algorithms where we can reverse engineer the price of the competition.
And with very good experimental design we measure the will to pay.
No current provider offers this at the same time.
Also, our competitors offer a software and a manual.
Not a solution.
(SLIDE)
Are we compliance with the FSA for a transparent price?
. YES. We work with segments, so each policyholder with the same characteristics will receive the same price.
. We are fair, but efficient. If you need to lower your price, we can say how much lower should your price quote be.
If you want to charge more, we can say the optimal price.
(SLIDE)
Now let’s talk about abuse.
We are running out of time, let me be efficient.
The current anti-fraud procedure is: look to the surroundings around a claim.
We look to the provider. The health provider. The car repair shop.
Imagine that a maternity hospital makes a lot of c-section. Each and everyone can have the proper patter and medical justification.
But the overall expense will be higher, but above all the digit distribution in the expense will be different.
Imagine that a provider needs clinical authorization for expenses above 2000 pounds.
Whenever a doctor wants to make this specific procedure, he will split the expenses in two invoices.
With our algorithm we will be able to detect this overcharge, since we will analyse the digits that compose each invoice.
And what we're going to see that we have more expenses started with the digit “1 " than we would expect
(SLIDE)
You see, there is a statistical law for the leading digits.
We know that the digit 1 will appear as a leading digit in 30% of the time. As the digit 9, for example, will appear less than 5% as a leading digit.
(SLIDE)
We can take this a step further and check if the digits follow the right distribution.
For instances in this example we check the first two digits of every invoice follow the proper distribution (is the red line).
And we can see that expenses that there are too many invoices that starts with “2 and a 5”
(SLIDE)
As a result, we have a 66% accuracy of finding overcharge on providers-
(SLIDE)
We are making money with these two products
We have credentials that we can share.
Came talk to us
Let’s run a pilot