Zenith Insurance gets better data insight and cuts fraud with SAS

Zenith Insurance's Jason Cabral explains how deploying SAS has improved efficiency at the insurance company

Zenith Insurance has moved away from spreadsheets to data analytics, using SAS to improve business performance and more easily identify fraud through the better use of data.

"We didn't have a lot of analytics, we did a lot of things on spreadsheets," Jason Cabral, group head of pricing and management information at Zenith Insurance and the Markerstudy Group, told Computing.

He described the company's previous use of data analytics as "virtually nothing, to be honest". This led to Zenith building a data warehouse with SAS analytical tools in order gain better insight into data.

"Basically we built the data warehouse to combine all the data. So now we have an enterprise data warehouse with self service for about 300 users. Also things like storage processes and reporting for the business," he said.

Cabral explained how the use of SAS data analysis tools has brought a number of benefits to Zenith, enabling the company to become more efficient at what it does.

"There are a few key benefits. One is that you have one view of the truth, you have one set of information that's traceable coming from the data warehouse," he said.

"So when people see figures, they don't spend time arguing about what data is correct, and if someone has a different figure from a different area everything is traceable."

He added that it also beings benefits when it comes to the new EU directive for codifying and harmonising insurance regulation, which comes into force next year.

"It's paramount for the Solvency II directive as well, which is the new European-wide insurance regulation coming at the end of next year that we have to prepare for," he said.

"Without having a proper data warehouse where everything is traceable and you have everything documented, it's very difficult to meet Solvency II requirements around data quality and the like."

With a relatively small pricing and management information team, deployment of SAS has enabled Zenith to do things that were impossible before, as well as providing employees with all the information they need to efficiently examine cases.

"It's economies of scale because we have a team of 15 people but they're supporting over 300 internal staff and they also link up with external parties. To be honest, we wouldn't be able to do that at all. It would be unmanageable if we were creating the database from scratch in order to send things out to the business," he said.

"It empowers the business, allowing them to create their own reports and refresh those daily and manage their own business area."

In addition to improving data insight, SAS has enabled Zenith to become better at spotting cases of fraud, with analytics speeding up the process for brokers and enabling the firm to provide a better service.

"Before, when a claim would occur, we needed to inform the broker. The department would literally have to send out a letter saying this claim has occurred on this policy. We used to have to manually send out letters and they could get lost in the post and the broker could say they never received them," Cabral explained, adding that his caused problems for the insurance firm and its customers.

"That can affect the customer's renewal for their policy, because their rate won't be adjusted for the fact they had a claim."

However, since automating the system using SAS, such mistakes are now a thing of the past, allowing Zenith to better detect cases of fraud.

"But that's all been automated using SAS. Now a lot of our interactions with the brokers around claims, no-claims bonuses, that sort of thing, has been automated directly to the broker and they also have access to some of our figures as well," said Cabral, who went onto describe how the deployment of SAS has made brokers more efficient at identifying fraud.

"In the fraud department, we have fraud investigators. It's basically made that department more efficient because in terms of the cases they'd identify we use SAS to build up models of which cases they should investigate, those that are more likely to be fraudulent," Cabral told Computing.

"Instead of reviewing 100 cases to find one case of fraud, they'll only have to review 10 to get that same case of fraud."