InsTech London’s Ali Smedley speaks to Jay Gentry, Head of Insurance at SafeGraph, about the company’s data: how it’s curated, how it can be used in insurance, and how machine learning plays a role.
Ali: What business problem is SafeGraph solving with data?
Jay: The leadership team at SafeGraph had previously founded LiveRamp, a marketing data platform company, and wanted to use their expertise in another way. They found that there were limited data sets about the physical world. SafeGraph was established to provide validated data on locations, or points of interest (POI) which are updated frequently and reflect what is actually present in the world. As the company grew, we realised that how people move around these places is also important, so we started to provide mobility data as well. Having trusted data in these areas is crucial to a variety of businesses, including insurance.
Ali: What does your role of Head of Insurance involve?
Jay: As Head of Insurance, my focus is to introduce the SafeGraph data portfolio to the property and casualty and health insurance marketplace. I worked at Pitney Bowes (now Precisely) for 15 years, so I have a long history of working within a data business. SafeGraph is a source of truth for the physical world, which makes us well positioned to provide data to the insurance industry.
Ali: What data sources does SafeGraph use?
Jay: SafeGraph’s machine learning and data curation process is a large part of our value. To obtain data on locations, we extract information from publicly available content. Today, we have 10 million POIs from around the world. In terms of the footprint of the physical place, we use multiple data sources including satellite imagery, and even do some hand drawing. For our mobility data we license aggregated and anonymised data that has been sourced from mobile applications where users have opted-in to sharing their location. We don’t manage individual device level data – all of our mobile data is anonymised.
Ali: How does SafeGraph use machine learning in its data collection and processing?
Jay: Machine learning is really important for making our business scalable. We provide data in over 88 countries and counting. Our algorithms ensure we can ingest and use large volumes of data. SafeGraph data is machine generated, human verified. Daily human validation is vital to ensure that the machine learning is effective.
Ali: Can you give some examples of insurance use cases for SafeGraph’s data?
Jay: Co-tenancy is an area that we’re seeing a lot of interest in. Co-tenancy is the ability for an insurer to know not only the location and the type of business that they’re underwriting, but also what other businesses and buildings are around it. Another area is the use of our movement of people data for commercial general liability insurance. Typically, insurers use revenue or square footage to underwrite risks such as slip, trip and fall. SafeGraph’s mobility data can provide insurers with information on how many people are moving in and out of a business, which is a much better proxy for evaluating the risk.
Ali: What other areas are SafeGraph looking to expand into?
Jay: SafeGraph’s core focus at the moment is in the commercial property space. Over the next few months, we’re going to be exploring how to apply our approach to the residential space. We want to maintain our anonymised data approach, so we’ll be investigating how to curate publicly sourced information for residential properties.
Ali: Why did SafeGraph join InsTech London as a corporate member?
Jay: We work with e2value, which has had a great experience with InsTech London and recommended membership to us. InsTech London gives us the opportunity to connect with companies and people that we wouldn’t have otherwise. We want to learn about potential opportunities for SafeGraph within the insurance industry. As part of our outreach I recently enjoyed speaking at the InsTech London live event: ‘Property Risk – What’s New and What Works?’.
Ali: What companies are you looking to connect with?
Jay: We want to connect with property and casualty underwriters and reinsurers. We’re also interested in exploring parametric insurance, so we’d like to connect with companies that we can learn from and potentially work with.
For more information on SafeGraph, go to safegraph.com