Meet the insurtech: Archipelago

Hemant Shah first founded RMS, a catastrophe-modeling platform, in 1989 as a graduate student at Stanford University. Born from a school project, RMS now serves as an industry leading financial risk modeling business for financial stakeholders to better understand the risk of insuring and lending money on their properties.

Shah notes, however, that as RMS continued to expand globally and evolve its modeling capabilities, he found that there was still room for improvement when it came to providing accurate data for model projections.

“Over the years, we kept developing these models for more perils, more geographies, but throughout the whole journey, the focus was on not just, ‘how do you quantify the risk?’ [but] how do you enable the insurance industry to better underwrite price, manage the risk, transfer the risk and hedge the risk at capital for the risk?” Shah adds. 

Archipelago

The drive to solve the issues he encountered through RMS – problems with securing accurate, high-quality data from which to forecast risk – led Shah to co-founding Archipelago in 2018.

“What I really learned,” says Shah, “was how the insurance industry works, how it creates value when it works well, why it doesn't work so well sometimes and then how to address those problems that led to the creation of Archipelago.”

Archipelago, a streamlined platform for commercial risk property data, deploys AI and machine learning to extract information from documents and unstructured sources about buildings found in large portfolios. The company focuses on large owners of real assets, such as real estate investment trusts or large corporate owners of property, that manage, maintain, engineer, lease and insure their properties. 

Archipelago’s platform allows these owners to obtain all the possible data and detailed information available about each property. This data – often found through varied sources, such as engineering reports, structural drawings, photographs, schematics and so on – enables property owners to better understand risks and makes the information available to insurers that can then underwrite the data. 

Shah notes that having an underwriter manually reviewing this information is often “a very cumbersome process.” Archipelago’s ML technology streamlines this by reading those documents and extracting high quality data about each property that is needed to ascertain its risks. 

This technology also recognizes patterns from the data it extracts. The platform identifies and organizes relationships between the data, allowing Archipelago to ensure that the information is accurate, detailed and connected within the database. 

“Machines are really good at seeing and extracting the patterns in a way that often humans would struggle to see,” Shah says. 

This information can show similarities between buildings or specific loss experience of a property, as well as ways to reduce risks found in similar data. 

Shah also mentioned that Archipelago is seeking to broaden its focus beyond digitizing just commercial real estate portfolios for property owners – he plans to expand across asset types and between property companies and operating companies. This includes amassing more data for the platform and seeking patterns. 

The company has already expanded to digitizing different data; so far, this information includes data for warehouses and logistics centers, apartment buildings, office buildings, shopping centers, hospitals, university campuses and data centers. Archipelago also plans to expand its footprint from owners of real estate to companies that have property insurance programs – those that may not own the property but insure the contents inside the buildings. 

“It’s a thrill being able to be at this early stage,” Shah adds, “innovating in a fundamental way to solve a really deep problem that cuts across the entire industry vertical.”