Intelligent automation and machine learning are key to handling the large amount of data that managing general agents have to deal with every day, as Mike Kelly of AutoRek tells Intelligent Insurer.
Dealing with data is a key challenge for any company. For insurance companies and managing general agents (MGAs), the amount of data that they have to be able to access is a truly formidable one—even for the smaller ones, as attendees in May at Intelligent Insurer’s Intelligent Automation & AI in Insurance conference heard.
“It’s an iterative process of learning business rules, which is very hard if you don’t have an intelligent system such as AutoRek to help you do it.”However, analysing and organising the data can be a challenge, according to Mike Kelly, insurance sector lead at AutoRek.
AutoRek has been around for 25 years and has more than 60 clients. AutoRek offers data management solutions that allow insurance firms to fulfil their financial, operational, risk control and regulatory reporting obligations. It automates the reconciliation processes for clients who have a large amount of financial data, such as premiums, claims, transactions, and commissions.
“We use machine learning and intelligent automation with these large datasets, where you’re trying to reconcile one set of data against another. Inevitably you can match up a certain number of items—and we’re talking millions here,” says Kelly.
“In the case of one of our larger clients they match 20 million records against another 20 million on a weekly basis.”
As a result of these large figures, when a company is going through that process, there are always a lot of exceptions in the form of data items that simply don’t match. According to Kelly, that can range from 10 percent not matching and in the worst-case scenarios, up to 70 percent not matching.
Kelly explains that AutoRek implements match rules based on business logic: as those rules are applied automatically, more matches occur. This leaves a residual amount of data, from transactions that don’t match, which will require human input in order to analyse them.
The AutoRek software learns to imitate those human actions. When a manual process starts to take place matching items based on certain categories of data, AutoRek analyses the data patterns being created and will suggest match rules that can then be automated.
As a result, Kelly claims, it’s possible to get down to just 1 or 2 percent of unmatched items.
“It makes the human involvement of matching items much easier as there are a lot less to look at. It gives you time to focus on the ones that need time to analyse. ” he says.
“It’s an iterative process of learning business rules, which is very hard if you don’t have an intelligent system such as AutoRek to help you do it.”
Kelly stresses that these automated systems are needed given the regulatory challenges that are approaching.
“For large insurance companies and for IFRS17—the International Financial Reporting Standard that was issued by the International Accounting Standards Board in 2017, which will replace IFRS-4 on accounting for insurance contracts in January 2021—accounting standards will change in the next few years,” he explains.
According to Kelly, these changes will require companies to look at their entire portfolio of business to see if they can report back to say what’s changed on a rolling basis for each of the areas. Companies with multiple systems for this data will find it very difficult to bring it all together and be able to say with confidence that they know that this transaction or this claim matches against that policy which was written decades ago.
“You need a system to bring all that together and to be able to reconcile and report back to the business,” he says.
A key challenge for insurers is Solvency II changes. “Further changes are being made to the Solvency II requirements, as well as the accounting standards changing for IFRS-17,” Kelly says.
“Those changes will require a serious review of the internal systems that each insurance company has. To change those systems and understand the data that resides within them and how you change those systems over the course of the next two to five years in order to adhere to these regulations is a big challenge.”
Mike Kelly is insurance sector lead at AutoRek. He can be contacted at: firstname.lastname@example.org
AutoRek, MGA, Artificial Intelligence, Intelligent Automation & AI in Insurance conference, Mike Kelly, UK