Tensorflight, a property data and insights provider has announced the launch of TensorAssist, its new chatbot assistant powered by OpenAI’s ChatGPT-4 technology. TensorAssist is available to all existing Tensorflight commercial property insurance customers. It allows them to rapidly understand the vast array of data points in its property intelligence platform.
With TensorAssist, customers can achieve a much deeper understanding of how Tensorflight evaluates risk in a fraction of the time. As a result, it will be able to perform more accurate and efficient analyses on properties on a much larger scale than ever before.
The Tensorflight platform includes property data garnered from three sources of imagery – ground level, aerial, and state-of-the-art satellite images – as well as data from public, contributory, and proprietary sources to provide comprehensive coverage of 99 per cent of properties in developed countries.
Tensorflight’s extensive property data platform includes information on building attributes such as construction type, roof pitch/geometry, and a number of stories. This is in addition to more detailed features like wall composition and condition, facade condition and even the presence of highly-combustible aluminium composite panels (ACP).
Designed specifically for the property insurance sector, Tensorflight’s virtual AI assistant will serve as an invaluable property analysis expert and will eventually be able to provide detailed information and insights on any property requested by the user, as well as offer guidance on associated risks.
Standardising unstructured text documents
Commenting on the launch, Jakub Dryjas, CEO, Tensorflight said: “With TensorAssist, our goal is to set a new benchmark for speed and accuracy in risk assessment in the commercial property insurance sector.
“Our AI assistant enables insurance underwriters to receive detailed information that previously relied upon the support of our team, transforming the entire process. The technology not only increases efficiencies but also allows us to streamline operations, reduce costs, and ultimately offer even greater value to our customers.”
Tensorflight is also harnessing advanced GPT text processing capabilities to address the challenge of standardising unstructured text documents, such as Statement of Values (SoVs) and building permits, which typically arrive in a wide variety of formats.
By leveraging this technology, Tensorflight will ensure the consistency and coherence of these documents, further streamlining the underwriting process. Additionally, Tensorflight is enhancing its property data underwriting initiatives by offering concise textual summaries of properties and building portfolios to provide further critical insights for insurers.
Looking ahead, Tensorflight’s longer term ambition is to use customer feedback from TensorAssist to build a fully-fledged AI underwriter that can operate with full autonomy and superior accuracy. This advanced system will be capable of receiving instructions and guidance from a senior (human) underwriter within an insurance company.
By combining the expertise of a senior underwriter with the capabilities of an AI underwriter, Tensorflight aims to empower a single individual to efficiently oversee the workload of an entire autonomous underwriting team, revolutionising the entire underwriting process while streamlining operations and maximising efficiencies.
