Tensorflight provides highly accurate and near-instant data about commercial properties. The innovation is based on Computer Vision and Artificial Intelligence, capable of analysing multiple image types and building a database of all buildings around the globe.
Current Status
We have signed a contract for the analysis of 250 000 commercial properties in the US for the next three years w Nephila Capital (they have become one of our strategic investors from the insurance industry). It's the result of the successful pilot project.
Moreover, we have recently finished multiple paid pilot projects and provided data on thousands of commercial properties located in the USA, the UK, and Poland. The results are satisfying for the insurers we are negotiating multiple contracts (the biggest one concerns the analysis of 450 000 addresses )- single analysis and contract for the analysis of 1.3M addresses annually.
In the past two months, we have signed 2 new contracts with companies in the top 5 of the global insurance market.
Winner of the GovTech prize in the contest for an artificial intelligence system that recognizes smuggled cigarettes on the X-ray imagery of trucks, rail cars, and other vehicles crossing the EU border.
Quarterfinalist of Zurich Innovation Challenge 2020 UK.
Featured in 2019 InsurTech 100, a list of 100 most innovative companies in insurtech around the world, created by FinTech Global
Laureate of Eagle of Innovation 2019 by Rzeczpospolita in the category “Best technological solution- project”.
Selected by Fintech Innovation Lab in New York as one of the top 6 startups in InsureTech.
Problem or Opportunity
The first problem that Tensorflight tackles is the inaccuracy/ non-existence of data about buildings around the world. That information is essential, for, e.g., better risk modeling, reinsurance, or underwriting. The second problem is price, latency, and scale of building inspections - sending a human inspector costs a significant amount of money, takes time, and is not efficient while an automated system does not face any of these problems.
Solution (product or service)
Tensorflight offers a unique solution, which is instant risk-related information about a building with global coverage, analyzed by our AI algorithms. To do so, we combine multiple data sources such as aerial, street view, and satellite imagery. We offer low cost, unique solutions, matching our clients' needs, and our results are over 90% accurate. We deliver information about the building features such as the number of stories, construction type, roof pitch; also information about the footprint, distance from objects and geolocation, and more.
Competitors
1) Verisk ProMetrix.
2) Cape Analytics
3) Recosight
4) Betterview Platform
5) EagleView
Analytics and measurement companies engaged in automatic property analysis. However, no other company on the market implements solutions as we do for commercial properties.
Advantages or differentiators
We differ in 3 key areas:
- Commercial properties: We are the only company focusing on commercial properties. There are many differences when dealing with commercial properties. For example, different features like occupancy type or various problems like the address to building mapping.
- Focus on the facade: In addition to overhead, nadir view, we are focusing on imagery of building facade that allows us to extract information like construction type of occupancy type.
- Deep learning: Our technical team is much stronger than our competition. Zbigniew is a co-author of one of the key computer vision publications in the last few years, creating the best classification model in the world that was the first to beat human accuracy, classifying 1 million images into 1000 classes with 96% accuracy and receiving 400 academic citations in 2 years. Both co-founders worked for over two years at Google, including teams like DeepMind, Google Brain, Google Maps, Search, or Google Compare.
Money will be spent on
R&D, marketing, hiring top talent in machine learning, sales efforts
