ThreatMark Clair is the SaaS platform for account opening fraud prevention in online consumer apps. Based on behavioral biometry and the analysis of granular data from online user sessions, it extends your risk engine and spots bad actors in real time.
Clair collects detailed anonymized data about the visitors of your online application, from their device parameters (device, system and network information) to behaviometrical patterns (mouse movements, keystroke dynamics, site navigation).
Clair creates a global profile of the target group of customers, using the collected data and state-of-the-art machine learning algorithms. It can then assign a score to each new user indicating how likely they belong to the target group. For instance, how likely they would be fraudsters or insolvent.
The score provided by Clair for each group is actionable. By default, Clair focuses on precision, but the settings can be easily modified according to specific business needs or marketing policies. It is also possible to create a completely new model and immediately test its value.
After calling our prediction service you will get all the issued scores in just a fraction of a second.
We use a JavaScript probe for web applications and a SDK for mobile apps. However, optional user information can be provided via API for elevated context and increased prediction precision.
Using our in-house developed algorithms, we can reliably detect multiple requests from the same device, or devices that have already been detected in suspicious activity.
By using behavioral biometry and collected authentic data, we can characterize the end user as a human being and relate him/her to any predefined group of customers.
Besides behavioral biometry, we use optional user-related data to provide deep analysis of e-mails, phone numbers and networks. Every score will be explicitly provided according to our WhiteBox principle.
The whole system is fully under controll
Full information for every loan entity, including scores for every model and device fingerprint
All available information for every user visit (even if user haven’t asked for loan), including visualization of his/her actions timeline
Seamlessly tell legitimate customers from fraudsters in real-time.
We tell you which are the most valuable users for your business.
Reduce manual processing and improve risk decision workflow.
Actionable scores with minimal false positive rate.
Shorter request-to-approval time, better business scaling.
Fully outsourced cloud solution with clean API.
Absolutely! We supply every new customer with a free trial. You will pay only if you decide that Clair has an added value for your business - after PoC we provide extensive report of what was done and how much money could be saved.
It depends on your loan operational volume, accounting rules in your country and in your company, what targets you set up for Clair (and if there are some customizations). On average a trial lasts from 2 to 4 months.
Yes! However while every new installation is providing added value from day-0, the full detection power will be achieved with all the behavioral biometry models. Some of them require additional training on data from your customers (which will be done during the trial period).
With our pseudonymized data collection strategy we cannot relate the data we collect to an actual person (unless we directly ask your operators). With disregard to this we treat all data we collect as PII and apply all the necessary steps to secure it, as well as fully support your customer rights (incl. "to be forgotten").