The responsibilities of a platform: Fraud Prevention
From the outside (believe me, I was there once) it looks simple. Starting a platform is perceived as “Just connect different parties online and be done with it”. I started Trade Cover Exchange early July this year and 5 months in I’ve learned about the implicit responsibilities of being/creating a platform. In these series of blogs I’ll write about what I’ve learned.
“Just connect different parties online and be done with it”
In this part I’ll talk about fraud prevention. I’ll tell you why I think this is the responsibility of a platform, why this is the case in our platform and how we will solve this.
First let me briefly tell you the goals of our platform. We want to make credit insurance easily and transparently available to corporates. Credit insurance has been around for a long time and insures businesses against non or late payment by their customers. Our platform brings insurers and businesses together and gives the opportunity to insure as “little” as one invoice (credit insurance before we came in was usually given on whole turnover only).
Why is fraud prevention the responsibility of the platform?
My thoughts before: “Insurers already do their own fraud prevention, why should we add this functionality as well?”.
Well, that might sound like a valid point. However we are opening up the credit insurance market to a far larger market. Higher volumes mean higher numbers of possible fraud. These higher volumes (on typically lower amounts than in the past) mean the fraud prevention instruments of insurers can’t cope as well anymore. We need to fix that for them.
When bringing more business (insurance requests) to insurers, a platform also needs to guarantee the quality of those requests.
Let’s give an example of possible fraud on our platform. This is not a real life example, just a hypothetical. Picture a guy working at a business and sending an invoice to a customer. They insure this invoice through our platform, but the customer turns out to be a shell company that he himself set up. That customer will not pay, the insurer pays the claim. And fraud guy is a few euros richer. Or picture sending out fake invoices and insure them. They will not be paid of course but fraud guy might claim this with the insurer.
In general, these cases will not occur. Insurers know perfectly well if a company or invoice is fraudulent. But, given the added higher volumes of everything, more of these cases might slip through the cracks.
How can and should a platform solve this?
A platform is the first “gate” for all parties involved. And since a platform is typically a digital only effort it can benefit from all data out there to prevent fraud. For a platform it is far easier to connect multiple sources of data and give a good overview of its customers and requests.
It’s easier for a platform to connect all the dots and determine if fraud is involved
Let me give you a few examples we came up with:
- Know our users, we check if their email address is theirs, if the domain name of that email address is owned by the company they say they’re from by scraping it from their websites.
- Know where our users log in from. A web session is checked for being fraudulent by making sure no bot is involved, have a second factor in authentication and we also check the location of the session to see if it is way off than usual login attempts.
- You can easily check if a user is actually a person. Services like Onfido make it easy to add verification and know your customer (KYC) functionality to your application.
- You can easily check if a company has connections with the company it tries to insure. Services like Creditsafe make it possible to search for companies globally and look up their officers, owners and beneficiaries. If there is a connection between the two you can red flag the request.
- Even more realtime are services like Owlin. Let’s say a user wants to insure a company because he knows that company will go under soon. The Owlin platform lets you read the market and detect risks far earlier than any human can. Check a company and its customer to spot negative sentiment and, if so, fraud might be involved.
- Ask for proof and have it checked by Machine Learning. In our case we can ask for an actual invoice or contract and have a machine learning model determine if the file has been tampered with or is out of the ordinary. Being part of the Google for Startups program, we have access to some great new technology for this.
The biggest reason of them all for actually pursuing and being great in fraud prevention: it is an awesome unique selling point for your platform. You will stand out as a platform that can cope with large volumes while still maintaining and guaranteeing the quality of business flowing through it.
Disclaimer: we have no financial interest or connection with any of the services and companies that I mentioned. We are a customer of some and would like to be a customer of all soon because of the services they provide
Want to stay informed on how we’re doing with fighting fraud? Sign up for our newsletter at https://tradecoverexchange.com
