Standardise data to leverage ‘wasted’ information, urges insurtech

When assessing claims information and accurately pricing risk, insurers trawl through immense amounts of data. But this valuable commodity often goes to waste once it has been stored in Excel spreadsheets in Lloyd’s and non-Lloyd’s formats, Alexis Renaudin, head of data science at insurtech Artificial, told Intelligent Insurer.

These spreadsheets are neither structured nor standardised but if insurers were able to manage and store the data in a more consistent format, the underwriting process could become far more efficient.

“We spoke to insurance clients about their major pain points and they told us...

...the amount of files they were receiving was overwhelming,” said Renaudin. He added that it wasn’t just the quantity of data that was causing insurers problems but the variation in the format of the data. What’s more, in the past few years the Lloyd’s market has been diversifying more into delegated markets, according to Renaudin, this increasing amount of data presents a real challenge for the industry.

According to Renaudin, Artificial help insurers in this market doing business through Lloyd's, focused on high volume and low premium claims. They recieve a large amount of claims reports in Excel formats from different brokers or third party claims administrators. He added that the quantity of these Excel files and variety of formats mean the data is generally not touched but rather archived or manually processed.

“Clients will receive thousands of reports on policies each month with multiple columns of data which they need to extract to better understand their risk. We analyse these Lloyd’s and non-Lloyd’s bordereau [report of claims] reports and identify relevant keywords and columns that we want to extract data from. For example, insurers might receive Excel files from over 50 different brokers in various formats and we then standardise this using data extraction tools,” said Renaudin.

On average,the insurtech claims to be able to extract 90 percent of key data points which it then consolidates into one master table. Artificial presents this data to their clients, providing them with key information on claims such as loss description, date of loss and total incurred loss. Insurers are then able to analyse this data using tools like Tableau to better assess their risk and improve underwriting decisions.

Artificial is an insurtech startup focused on creating a machine learning enabled insurance platform. Having received $5.2 million in funding since its inception in 2013, the insurtech is now set to begin its Series A funding. It also plans to unveil the insurers it is working with in the coming weeks.