Client Description

Trōv is the world’s leading On-Demand Insurance platform, intelligently protecting anything, anytime, anywhere. Trov’s consumer application enables people to insure single items, for just the period of time they need, entirely from their mobile device. Trov also provides tailored insurance technology for companies innovating in the mobility space. Founded in 2012, Trov is available in Australia, UK, and the US.

Invested Amount

$200,000

Status of Engagement

2016 − ongoing

Team Involved

6 people

Technologies

Machine learning algorithms, Python, NLP: errors detection, sentiment analysis, Clustering techniques: LDA, K-means, DBSCAN, t-SNE Neural Networks

Service performed

Software development, QA, PM

Business Challenge

The Client required robust system based on diversified metrics to deliver immediate fact-based insights utilizing a vast amount of enterprise data. The team also had to conform to the security regulations and data protection laws.
Working with real-life information which includes private and sensitive user data was made lawful as our team is GDPR compliant and conform to regulations.

Trov mockup

CoreValue’s Role and Provided Solutions

The client aimed at identification fraudsters (suspicious clients) before claims are paid. Based on Artificial Intelligence and statistical techniques our data science team built a robust fraud detection system applying:

  1. Machine learning techniques to automatically identify characteristics of fraud
  2. Neural networks can learn suspicious patterns from samples and are used later to detect them
  3. Accumulated vital metrics combined with the utilization of 3d party services that signal the potentially suspicious clients

The system enables the client to make fact-based decisions through providing access to enterprise data for easy analysis and also:

  • Automatic fraud detection
  • Improved customer experience
  • Lower loss ratio
  • Accelerated payments on genuine claims

What’s Next

We continue working on additional data sources for analysis. We do expect to extend collaboration with the client that will include machine learning practices and other data science algorithms in order to provide deep insights on customer data for continuing business success.

“From a data science perspective, the people from Core Value involved in the work have had tremendous success. They’ve helped us detect and reduce fraud by a large portion, by providing key metrics in that direction. They’ve been able to improve our product and ad spend significantly, informing our team on our marketing targets, and how to improve efficiency.”

Mark Merhom

Data & Analytics Engineer,Trōv

title