Vlucht-vertraagd.nl case

Machine Learning Claim Engine

 

Introduction

Since 2011, vlucht-vertraagd.nl has been helping airline passengers successfully file claims for compensations for delayed, cancelled and overbooked flights. Their mission states that every passenger will be able to take full advantage of their given rights as airline passengers. This requires the difficult legal process to be made easy to understand, transparent and efficient. Every single day, 500 to 1000 new claims are filed at vlucht-vertraagd.nl.

The vlucht-vertraagd process

Claims at vlucht-vertraagd follow a pre-set process:

  1. Calculation: what is the probability that a claim will be successful?
  2. Validation: is the information provided correct?
  3. Completion of profile: once assessed, all necessary information needs to be collected in a clever and automated way,
  4. Assessment: if the probability that a claim will be successful is high enough, an assessment is made to determine if judicial process is worthwhile,
  5. Judicial process: go to court, to get your money.

During the these steps, in-house experts utilize internal and external information to determine whether the claim can move onto the next step in the process. This means copious amounts of non-automated, manual labor to validate a claim’s viability and chance of success. During peak season, summer time, to-be processed claims exceeds working capacity, causing a growing backlog.

 

Where does ML help?

In the past eight years, vlucht-vertraagd.nl has expanded significantly which helped establish a rich and ever-growing dataset containing both claims and flight information. This data set will be used to speed up aforementioned process steps by (partial) automation of the validation step by applying the power of machine learning. ML comes back in the following steps:

  1. Calculation: based on the information provided, and some external sources, can we predict the chance that a claim will be paid to the client?
  2. Validate: can we automatically detect anomalies in the application?
  3. Completion of profile: all kind of information is collected. Which information is most useful, what has most predictive power and should be prioritized?
  4. Assessment: a new calculation determines with high certainty if the claim will stand in court.

In this way, ML helps easing a difficult judicial process.

Want to know how we can setup your data science infrastructure?

Leave your e-mail address and we will contact you.

Case harvest

Enjins logo
Close

Contact

De Entree 234
1101 EE Amsterdam

KvK: 71755101

We would like to hear from you

Contact