Using its own constellation of microsatellites, ICEYE is providing insurers with high-resolution, building-level flood depths around the world immediately after an event.
VP of Solutions Charles Blanchet joins Matthew to discuss how the company is making satellite technology commercially viable, and how insurers are now able to provide rapid flood loss estimates and efficient claims settlements.
Talking points include:
- How synthetic aperture radar can see through clouds
- Why the insurance division of this company has been engaged by leading insurers in only 12 months
- How aerial imagery is augmented with third-party data
- Parametric product solutions
- Making a complex offering easy to use
Look out for ICEYE on our Live Chat panel on Thursday 1 April. The company will also be one of more than 60 featured in our forthcoming Location Intelligence report.
If you like what you’re hearing, please leave us a review on whichever platform you use, or contact Matthew Grant on LinkedIn.
Sign up to our newsletter for a fresh view on the world every Wednesday morning.
Continuing Professional Development – Learning Objectives
InsTech London is accredited by The Chartered Insurance Institute (CII). By listening to an InsTech London podcast, or reading the accompanying transcript, you can claim up to 0.5 CPD hours towards the CII member CPD scheme.
- Claim 0.5 hours for listening to Episode 128 of the InsTech London Podcast
Your satellite will be with you shortly: Rapid flood loss assessment – Episode 128 highlights
Matthew: Can you start with some background on where the company came from and what ICEYE is doing?
Charles: The company came from our two co-founders meeting in college and taking on a project that they were told was impossible. That was miniaturising synthetic aperture radar, which is technology that can observe earth through clouds at night and detect very small changes. ICEYE now deploys the world’s largest and highest quality, synthetic aperture radar constellation. We manufacture the parts, assemble the satellites, launch them into space, then manage the constellation and pull data down.
My team democratises the data, which is difficult to work with for non-specialists. Insurers shouldn’t have to deal with what we deal with to get it into a usable form, so we’ve made investments in third-party data and analytical technologies like machine learning and algorithms. We use those to produce data that can be directly used in insurance organisations systems and workflow.
Matthew: What is the story behind the name ICEYE? I believe it relates to the original application?
Charles: Our first round of funding came from Exxon, the oil company. The price of oil was very high and we were asked to build a technology that could help oil exploration on the polar caps. The oil price then dropped and we had to pivot and find a different market, which is why we ended up here.
Matthew: How long has ICEYE been building satellites?
Charles: The company’s been around for 10 years and there’s been a lot of research and development involved. The first test version of our technology was a radar on a university cart, the second iteration was on a plane, and we had to put prototypes into space before launching a working satellite.
We’re now at seven satellites in orbit and are starting to unlock use cases that require higher temporal resolution, which means we have to be able to get to places fast enough to be useful.
Matthew: How long does it take from a flood event happening to you being able to get a satellite to review it?
Charles: Capturing the high water mark of a flood is very difficult. It can be instant and wane very quickly or last for a while. How fast we get there depends on where it is on the planet. The further north or south the location is from the equator, the faster we can get there.
When I joined last year I did a comprehensive study of floods and concluded that we needed seven satellites to effectively do this. We had a successful flood season last year with just three satellites, now we’ve got a slightly easier job. The good news is that we’re on target to have 14 in orbit, doubling the capacity required to rapidly capture all of the floods that insurance cares about.
Because we had to start with not enough satellites, we had to get creative and add other data sets such as river gauges to extrapolate the high water mark. Now, if we get there an hour before and after, we have the methods and data sources to engineer that high water mark.
Matthew: How have you managed to miniaturise the deployment of satellites?
Charles: These satellites used to cost hundreds of millions of dollars to build, meaning only well-funded government programmes could afford to put them in space. There were papers written about how it was never going to be possible to make them smaller.
One of our mottos is ‘make the impossible possible’ and that’s what we did by miniaturising them. We made them a lot less expensive, making it commercially viable to put enough up in orbit to start unlocking difficult use cases, like capturing the high water marks of floods. In the past, putting 14 satellites in orbit would have cost billions of dollars. We’ve made it a millions of dollars thing.
Matthew: Can you explain how synthetic aperture data is different from more traditional optical satellites?
Charles: People are most familiar with optical imaging, which is great and we use optical imagery. It’s like the camera in an iPhone taking pictures of the planet, and it’s really good for use cases like agriculture where someone needs to see the colour of a plant to determine its health.
The problem with optical imagery is it’s not dependable. If there’s cloud cover or it’s at night, it’s not usable. Synthetic aperture radar can see through clouds and at night, and we can detect millimetre level changes. We can see a footprint in the snow, for example. Instead of a camera, we beam energy onto the planet, and that energy reflects back and we receive a signature that allows us to see things like water.
Matthew: There are two ways this could be used. One is for insurance companies that need to understand what the loss could be very quickly, and the second one would be the deployment of people to do the loss adjustments and triage. Are those the two main use cases for your clients?
Charles: Yes. We can help people understand loss numbers. Storm Christoph, which we just monitored in the UK, covered 20,000 square kilometres of land and we produced analysis within 12 hours of the peak.
We don’t do damage data; we do hazard data. We provide the depth and footprint of the flood down at building level accuracy. Insurers can take that data, convert it to damage and then convert it to losses using the same tools they already use. It allows them to understand the full loss of the event at the portfolio level. Then they can then set aside the capital without having to over or under commit, which is important to the core of the insurance industry.
Matthew: You’re providing that data as spreadsheets or CSV files. Do the companies you work with prefer to get it in that raw data format?
Charles: We made a firm decision to be a truth company, not a religion company. We’re just observing what happened and giving the insurance industry and governments that reality. We’re not trying to get into the business of understanding damage and converting that to loss. There are thousands of companies out there that have that core competency. We stick to what we’re really good at, which is observing what happened.
Matthew: You mentioned complementing aerial imagery with flood depth gauges. How are you combining different sets of data?
Charles: Third-party data is a huge endeavour for us. We have a team dedicated to acquiring and onboarding third party data, with a multi-million dollar budget assigned to it. We found that synthetic aperture radar data is key, but without a good digital elevation model, great weather data, great building footprints, we can’t deliver the degree of accuracy that our clients are demanding about water depth. It’s a big data cocktail that we have to mix to make this possible.
Matthew: Are there specific sources and companies that you’re partnering with?
Charles: There’s a lot of free data that we acquire. Other data we have to purchase commercially and we have about 20 commercial data vendors that supply everything from digital elevation models to building footprints and weather data.
Our technology can also produce some of this data, so we can build some of the datasets using our satellites.
Matthew: You’re bringing in major clients including Tokio Marine, Swiss Re and the World Bank. Can you talk about what you’re doing with those organisations?
Charles: We’re developing a full complement of data and technologies with the World Bank and they want to understand when a flood is going to happen. Finding potential floods is a huge problem that we’ve had to build an entire set of products around. We will probably be investing $5 million in research and development this year on that alone. They’ve created an entire ecosystem, for us to capture the flood, and they’re going to help us measure the performance of it.
We’re also looking to create a parametric offering for the government of a country in East Asia, so we’re doing a lot of research and development around that. We are helping them enable the automation of their claims process. We know that in certain locations, where we are very confident in the water depth measurements and the evidence of loss is so reliable, that insurers can automate some of their claims, triggering immediate payments to their customers.
Matthew: We’ve seen a lot of floods in the last few months. How has what you’re offering played out in terms of practical validation of the loss estimates or the depth estimates?
Charles: The interesting thing we’ve learned is that there are so many floods. We are assessing six globally right now whilst you are talking to me. I’ve never seen the daily number go below three and the most we’ve seen is nine.
Showtime for us is the typhoon, hurricane, tropical storm season that’s coming in the next several months. We’re running practice drills and we’re putting more satellites in orbit between now and then, so we’ll be ready by the time the June/July season comes.
Matthew: Hurricanes can carry a lot of water with them as rain, or they can drive a storm surge or a combination of both. How do you make sure you’re ready to give the data to clients once a big storm appears?
Charles: Operationalising the use of this data is a big change for our clients, so we’re developing an advisory practice that helps people understand the business rules that need to be created to take full advantage of this data as it comes in. The technical implementation is very simple. It’s the business change in how that information is used that’s important.
We have several operations rooms: one for the satellites and one for tracking the floods. Then we have a flood team, including a full-time meteorologist, that does this all day, every day, and we’re creating purpose-built internal software systems. There are different ways that we work with the constellation of our satellites, depending on the severity of the situation and how important it’s going to be for insurance companies and governments.
It’s amazing how fast we’re moving. We’re talking to our clients about concepts that didn’t exist in our business 90 days ago. In that time we’ve assigned five dedicated people to new solutions, we’re building software products, and drawing in tonnes of data to make consistent decisions around what we go after.
Matthew: Are insurance companies able to get access to your forecasting data, or at least alerts related to floods around the world?
Charles: We’re including that in our regular products. We’ve built that concept for floods and now we’re looking at engaging with a big loss adjuster company. They find that the alert data is just as valuable as the flood data so they can mobilise resources.
Part of our onboarding process is proposing that they take the two different data sets: the alerting data set and the flood hazard data. Then we figure out the systems our clients need to integrate with to operationalise the data and the business processes that are required to deploy people and do what they do.
Matthew: A significant percentage of loss can be saved if companies or individuals can take action before a flood occurs, whether it’s moving possessions upstairs or putting flood defences in place.
Charles: Proactivity is something that some of our clients are really interested in. We’re building a product for one large insurer that we’re calling persistent insured asset monitoring. What we’re going to be doing is imaging the insured assets every month and delivering a simple report to the client that says, ‘here are the assets that have changed’. Then they will go out into the field, investigate those changes, and bring that data back to us.
What the insurer wants to prevent is water going through a hole in the roof and getting into the electrical systems and causing a fire, then the whole building is a loss. They want to be proactively telling their client that something is wrong with their building and help them fix it before the problem happens.
Matthew: You’re starting to look at other hazards. Can you talk about what might be next for ICEYE?
Charles: We’re going to use that persistent insured asset monitoring solution as a baseline to start feeding in field reports. A flood is different from an absorbing water problem, but when you look at wildfire, wind, earthquake, all of those are change detection problems.
We’ve grouped all those together and begun a research and development initiative in partnership with several clients to get into the damage layer. Imagine us being able to say whether or not a building has been impacted by a wildfire and to what degree. When I’ve had conversations with FEMA, that is really important to them. Twenty-four hours after a fire, they want to deploy cheques. With our data, we’re going to be able to do that for wind and earthquake as well.
Matthew: When you talk to people in insurance organisations, who are you finding is the most engaged?
Charles: It’s organisations with leadership that have a firm commitment to innovation and transformation. No other company has been able to offer what we are offering today. We’ve found we need to be talking at senior levels in the organisation, in the C suite, because this is a board-level decision to transform a business and may need to be funded outside of existing budgets.
We find that everybody in the organisation is interested and fascinated and wants to make these changes, but it takes a firm commitment from senior leadership to make it happen.
Matthew: You’ve very generously supported us at InsTech London as a corporate member. Why did ICEYE decide to join?
Charles: We have something really special and we want to let everybody know that it’s available. We want to supply a steady stream of evidence about what we’re doing and InsTech London has the credibility and network to support that.
Ultimately, our goal is to be loud and proud about what we’re doing and get as many people involved as possible. The more clients we get increases that feedback loop between the field observations and our products. Every one of those helps us produce better, more accurate products.