Despite the adoption of digital technology, insurance remains one of the most traditional sectors in finance. Its continued use of legacy systems has hampered its development. However, organisations like Tensorflight are now doing more to ensure the industry starts to catch up.
Tensorflight provides structured property data for the insurance industry by instantly analysing aerial, satellite and street view imagery. Robert Kozikowski is the company’s co-founder and CPO; he has over a decade’s worth of experience in the tech field and is now using his expertise to accelerate the digitisation of insurance.
We sat down with Kozikowski to learn more about the company:
Tell us more about your company and its offering
Tensorflight is an AI imaging-based insurtech automating commercial property inspections and claims processing. Our mission is to make high-quality data available to the insurance industry to enable informed, timely and accurate decisions.
We use three sources of imagery – satellite, aerial and ground level – to automate the slow and expensive process of in-person property inspections. We also help insurance companies evaluate commercial and residential properties and risk as part of the underwriting process. Finally, we help them to better understand their portfolio exposures.
We are a Polish/US insurtech with offices in Warsaw, New York and London. We’re proud to work with many of the largest global commercial property insurers including Zurich Insurance Group and QBE.
What problem was your company set up to solve?
In-person property inspections are time-consuming and expensive. The commercial property industry is currently poorly served and has to make decisions based on inaccurate property infrastructure data. These are the problems we’re solving.
As mentioned, we use three sources of imagery – satellite, aerial and ground level – as well as data from public, contributory and proprietary sources and we can provide detailed coverage of 99 per cent of properties in developed countries.
Our solutions automatically assess a range of property attributes: building footprints, construction type, roof pitch and geometry, number of stories, and can even detect the presence of ACP panels – the layer of highly-combustible plastic identified as the primary cause of the rapid-fire spread at Grenfell Tower.
With our AI and computer vision technology we’re able to determine what a human assessor can perform looking at the same imagery – albeit on a much greater scale and without needing to visit the property. With our solutions, clients can significantly decrease both the cost of, and time spent on inspections in the insurance process. On average the cost of in-person property inspections can be reduced by over a quarter, and our solutions can considerably enhance the efficiency of the claims function.
Since launch, how has your company evolved?
Our company was founded in 2016, with a very scientific, engineering mindset. We were a team of PhDs and ex-Google engineers looking to apply our expertise in deep learning. The first breakthrough happened in 2019, when Nephila Capital, the very first company from the insurance industry, invested in our product. We’ve continued investing in research and engineering, gradually developing the product but also creating the company that we are today.
Our Series A fundraise led by QBE Ventures in October last year was a huge milestone for us. Since then we’ve signed a number of strategic partnerships which have enabled us to further enhance our offering. This year, for instance, we partnered with the Intelligence Business of Airbus Defence and Space, a global leader in geospatial data, intelligence and defence solutions. This allowed us to ramp up our property analysis and assessment capabilities.
The partnership means we now have access to the highest commercial resolution imagery available on the market. We also benefit from data with the best geolocation accuracy. The state-of-the-art satellite imagery will enable us to gather building-related information and detect changes to properties.
This ranges from information about roof condition and asbestos detection, to damage analysis and appraisal following natural disasters, such as hurricanes, and within conflict zones where on-the-ground inspections are impossible and other sources of imaging are unable to provide the level of detail required for the insurance process.
We’ve also recently made some notable senior appointments and hires. Jakub Dryjas, one of our longest serving members and our former chief of growth responsible for spearheading the company’s expansion across our key markets of the US, UK, Europe and Australia was promoted to CEO in the summer. We’re excited for the next chapter in our business under his leadership.
What has been the biggest challenge or most ‘tricky moment’ to overcome?
Digital transformation in the insurance industry is by far one of the biggest challenges. It is still very much in the infant stages and legacy systems continue to impede progress. The pandemic emphasised the need for the property insurance sector to adopt technological innovation, to change how the property inspection and insurance process is managed, and to reduce friction in the customer experience.
As a result, we are seeing a shift in mindset with insurers increasingly seeking out faster and more streamlined alternatives and beginning to embrace emerging technologies such as AI and machine learning which offer richer data insights and faster and more efficient services.
What are your biggest achievements or ‘proudest moment’ so far?
We are very proud of our pro-bono work with the Kyiv School of Economics (KSE) Institute providing and analysing satellite imagery to assess the damage caused to infrastructure within Bucha, Irpin and Mariupol.
Since the onset of the Ukraine war in February 2022, the KSE Institute, a leading think tank and the analytical unit of the Kyiv School of Economics, has been spearheading an initiative backed by the Ukrainian government. The initiative analyses information on the damage inflicted to civilian infrastructure and is effectively an open-source data project to evaluate and track the cost of war.
Through analysing satellite imagery and rooftop geocoding information, we’ve been offering insights into the extent of the damage caused to a particular building in the analysed areas, while also providing the exact geographic coordinates of the buildings. Our team of AI experts trained a deep neural network model to detect and classify war-damaged buildings on satellite images and to date almost 70,000 buildings in Mariupol, Irpin, and Bucha, have been identified and analysed using our model.
The project is also helping the Ukrainian government assess the reconstruction requirements of the country. For example, the intelligence will help inform whether to prioritise the rebuilding of a power station to prepare for the winter months, or a hospital or school based on the needs of residents in affected areas.
Our next plan is to analyse satellite imagery covering Lysychansk, Melitopol, and Popasna. However, satellite imagery is extremely costly. We are facing upwards of €5 million to scan the 100,000 square kilometres of Ukraine still awaiting damage assessment. Additionally, this data becomes out-of-date quickly when further damage is inflicted, so we are currently looking for sponsors and partners to help finance and continue the project.
How would you describe the culture of your company?
At Tensorflight we have an enterprising culture and our company is a meritocracy which values creativity, ideas and innovation. We believe this approach enables us to stay on the cutting edge and helps to reveal the best ideas. It also keeps our staff motivated and engaged while also fostering leadership.
As a company, we’re informal and we have a “people-first” approach. We also have a flat structure where everyone in the team has the right to express their opinions. They are encouraged to share them openly and often. And from the insurance sector, we have adopted a focus on reliability, security, consistency, and professionalism.
What’s in store for the future?
We have big plans ahead. We’re investing even more in product development as well as in developing our people, hiring the best talent and expanding our global reach. In terms of product development, we’re currently working on improving the accuracy of a number of key attributes that we assess, including geocoding and square footage.
Our plan is to enable digital insurance for commercial underwriting and make residential digital insurance more efficient and cost-effective. We’re working to enable all of the critical pieces insurers need.
We’re also gearing up for a Series B fundraise next year. This will allow us to further expand and enhance our offering.
