MEET OUR CODE_N CONTEST FINALISTS 2018: Tensor Flight from the USA
The way forward is to go digital. CODE_n CONTEST Finalist Tensor Flight has found the perfect solution to automate real estate inspections using computer vision and artificial intelligence.Learn how their technology delivers major value for the insurance and reinsurance industry in our interview with chief scientific officer (CSO) Zbigniew Wojna.
Jessica: What is TensorFlight all about?
Zbigniew: TensorFlight is all about great technology that delivers major value for the insurance and reinsurance industry. By drawing on the most recent innovations in the fields of computer vision, AI, and machine learning, we have developed a product that is without parallel in the market, also offering enormous business potential. Our software is a global innovation that completely changes how auditors inspect real estate. It automates time-consuming and expensive inspections otherwise carried out in person, simultaneously minimizing costs. Our solution was developed after a team of highly skilled individuals turned their backs on leading Silicon Valley companies to invest their energy in an amazing vision: TensorFlight.
Jessica: How did you come up with the idea?
Zbigniew: About two and a half years ago, after many conversations about starting our own business rather than working for somebody else, I made a decision with my long-time university friend from Warsaw – Robert Kozikowski (now the company CEO) – and we joined forces to set up TensorFlight. Our initial idea revolved around the agriculture and forestry industry, but we ascertained that these sectors of the economy did not offer sufficient potential for a strong AI startup. Soon afterwards, we determined that the insurance industry is open to innovation and has a specific need for which we can provide a product. Our solution automates real estate inspections using computer vision.
Jessica: What are you trying to solve?
Zbigniew: It is estimated that the insurance industry has lost a staggering $4.5 billion simply by mis-categorizing construction classes and fire protection measures. A problem lies in the traditional method of inspecting real estate. Information is outdated, and it can cost a lot of money to carry out inspections in a process that takes a long time and has shortcomings on a number of levels. Most insurance and reinsurance companies still carry out inspections in person. The average property inspection costs around $500. This is where our software helps. By using freely available images (satellite, aerial, Street View) and combining these with machine learning models, our solution can correctly assign values to individual properties. This has the potential to cut costs to as low as $3 for single inspections, which can also be carried out more quickly.
Jessica: TensorFlight is building a database for use in the real estate insurance market. Have you been thinking about reaching out to other industries in the future or adding other functions to your service?
Zbigniew: Our product features are best suited to the insurance industry, both in terms of the technology underlying our system and other processes. The strength of demand in this field of business makes us certain that this is the right way to take things, so we do not intend to change the technology for the foreseeable future. Of course, we are always keeping our minds open to reaching out to other industries. There could be strong potential in a number of areas, but this would require a major pivot for our startup, both in terms of the technology and our whole organization (our team currently consists of 15 people). That’s not to say we are unwilling to meet people and discuss different ways to use our technology, but for now our target is very clear – insurance.
Jessica: Thank you for the interview, Zbigniew!