You are here
Insurance 

The Co-operators is intensifying its efforts analytical

by

Alain Thériault

February 9, 2018 09:45

Carl Lambert

The Co-operators has put full gas on its team dedicated to the analysis of massive data and modelling, bringing its workforce to 75 people in 2017 to 110 today. The user experience, and data governance are at the heart of the emphasis placed by the Co-operators on its business analytics.

In 2017, the team of analysis of big data has helped to Co-operators to model a insurance that is able to take on the risk of flooding of river banks, normally uninsurable in the centre of Canada, despite the lack of historical data. She comes to see its efforts magnified and refocused, both in damage insurance and insurance of persons.

The team is in transition as well as a centre of expertise, who will take care of all the predictive models, which do not fall within the scope of the department of actuarial science. It is, for example, to determine the analysis of the surveys the factors that influence the satisfaction of customers, or trace the best routes to follow to increase the effectiveness of the inspectors of damage.

Growth of users

In the wake of a significant growth of the team in recent months, the Co-operators has expanded several functions. Responsible of the team, the national vice-president, business intelligence for the insurance of individuals and businesses, Carl Lambert, has seen its position move to vice-president, business intelligence, head of data and analytics.

Function he will be heading now has five senior managers, since a new one is added. This is Jason Smith, previously the deputy director, who takes charge of the part of the user experience. The rapid growth of this component explains the promotion of Mr. Smith.

«There were 2 000 logins per month by January 2017 and 7,000 in December 2017. We want to increase this number to 15 000, and that of active users, currently 1000, to 2500 by the end of 2018, » revealed Mr. Lambert, in an interview with the Journal of insurance. The other senior managers of the team are Clément Brunet, Simon Castonguay, François Godbout and Dominic Fortin.

Improve the experience

The Co-operators wants to ensure the continuation of this growth in measuring and improving the degree of satisfaction of employees with respect to the interface. It allows them to use the data of other employees and of clients, for example, in the framework of their administrative functions or human resources. Users can use this platform to generate reports and view data in self-service.

To improve it, the team uses the science of ergonomics, to a layout of the elements on the screen that is more efficient and easy to the eye. It also plans to increase the update frequency of the reports generated for the users.

Who can see what

Mr. Lambert states that the governance of data is not processed by the Co-operators of a binding. «We look at what we can do in governance, in order to give people what they need, without infringing the laws or regulations on the confidentiality of the data,» he says.

The insurer still wants to draw clearly the limits of data-sharing of the company. «It is also to ensure the quality of data and their treatment, to document the metrics, and to determine who has the right to see what» says Mr. Lambert. It will as well determine that may see an employee who uses data for the human resources services. Another could not see the contact details of bank payment to a customer, or address.

Sort the e-mails

The analytical part of the artificial intelligence and the science of measurement also fall under the umbrella of Carl Lambert and his team. She has designed, among others, an algorithm which sorts the e-mails of customers.

«A client who writes for a quote for auto insurance and another who writes to get his pink card (proof of insurance) are to be classified in their respective category in 88% of cases, there given in the example. In regards to the choice of metrics (measurability), we will choose which unit of measurement we choose, according to what we intend to measure. «

Related posts

Leave a Comment

Яндекс.Метрика