Using data science to prepare for GDPR regulation
Nicolas Meric, CEO of InsurTech and French AI start-up DreamQuark, explains how AI can help insurers and banks, and discusses DreamQuark’s business strategy and the new GDPR regulation.
Life Insurance International (LII):
What recent projects has DreamQuark been involved in that have led to insurance innovation?
Nicolas Meric (NM): To answer that question, I will start by explaining what DreamQuark is working on. We have built a platform which is very simple to use. With this platform, the aim is to enable insurance companies and banks to add value to all the data they have internally.
Our R&D team continuously develops advanced algorithms and deep learning technology, and we have made significant progress in making the algorithms more transparent.
With GDPR regulation coming into force on 25 May 2018 in Europe, it will be necessary for insurance companies and banks to be able to explain to their policyholders and customers why and how algorithms take some decisions. Our technology allows insurers and bankers to understand data, to clean data and prepare it for other applications.
At the end of the process of using DreamQuark’s platform, insurers have implemented really powerful and innovative fraud detection systems, anti-money laundering systems and also asset management tools based on their existing data.
How will DreamQuark’s technology help insurers and banks deal with the GDPR data regulation?
NM: Data is a key asset for these companies and a lot of information within the datasets is exploitable to address the new issues the insurance sector is facing.
The first step in using machine learning is being able to deal with data quality. On risks topics in particular, where a lot of information is available, using quality data is key. With DreamQuark’s platform, we automate this cleaning and refining phase on the data for all users.
GDPR brings a new light on the matter, as the regulation will require insurers to explain what information they put into the models they created with their client’s data.
With our platform, we help insurers be compliant by showing and keeping record of all changes made to the data; and by giving an excellent understanding of the model’s inputs and outputs. It allows insurers and banks to explain why an action was taken or a decision made.
From dealing with banks and insurers, how prepared do they seem to be for GDPR?
NM: it appears some players may not be ready for GDPR, but the situation was similar for the introduction of Solvency II.
France has a legal framework forcing companies using customer and employee data to provide information about how they will use the data. They also have the obligation to explain what type of data they gather and its purpose.
GDPR is going one step further into the protection of personal data in the sense that all companies will be compelled to follow strict guidelines as to how long and how they can exploit, or even possess, this personal data. This is applicable to data of all types, including unstructured data such as voice and images.
The purpose of GDPR is to write into law what companies need to do with all their data, identify who is responsible for the data, [and note] how they can work with providers of data mining technology and on the obligation from the provider in terms of protection and security.
All banks and insurance companies are racing to be prepared for the start of the GDPR regulation on May the 25th. The incentive is all the more important that companies that are not compliant may be fined up to 4% of their revenues.
A key for companies to prepare for GDPR is to hire the right DPO (data protection officer)
How has DreamQuark performed as a business over the last year?
NM: We have had steady growth over time and now there are 20 people in the team, compared to 11 people in October 2016.
The team is growing, and there has been several senior management recruitment over the past few months. We for example recently hired a chief financial officer. This is an important step and a lot of work has been done to improve the company’s structure and move from a “start-up in a garage” mode to a more structured company that can address our customers’ challenges.
By the end of this year, the company will have reached approximately 25 people.
We work with the UK teams of big banks, we work in Switzerland, we work in Africa, and our pipeline is growing very quickly.
What is DreamQuark’s business strategy for 2018 and 2019?
NM: The idea behind DreamQuark is to have a strong focus on financial services. We will work to move our product from a generic platform to specialised artificial intelligence applications for banks and insurance companies.
Now we want to put some templates into the platform that will further ease even the use of our technology for anti-money laundering, combating fraud and asset management.
We will work to make DreamQuark more automated, simpler and aligned with the needs of larger insurers and asset managers.
DreamQuark is a start-up born from the experience of its founder in analyzing Big-Data in particle physics. For more information contact: www.dreamquark.com