Spookfish AutoAssess uses the latest computer vision technology to predict vehicle repair costs using photos of the damaged vehicle. AutoAssess fast tracks the claims process, reduces the cost per claim and helps to mitigate fraud.
Spookfish started as a consulting company, providing deep learning advice to businesses, ranging from law firms to telecoms companies. This consulting period formed the basis of our market research, where we identified the auto-insurance industry as a scalable opportunity for a product powered by deep learning technology.
In addition to consulting, Spookfish continued to engineer deep learning models in both computer vision and natural language processing. Over time, the team began focusing on computer vision, building high-quality classification and detection models. These models form the technological foundation of AutoAssess, a Spookfish product.
Auto-insurance is a competitive industry in South Africa, where innovation that speeds up claims processing and reduces costs to the insurer can provide a significant competitive advantage. AutoAssess provides this competitive advantage to insurers by utilising the latest computer vision technology to speed up claims processing, while reducing costs and fraud.
After graduating from UCT, Robert spent two years at Eskom as a system engineer at one of the power stations. During his time there, he self-studied and successfully completed CFA level 1 to better his understanding of finance. This caught the eye of his then manager, who sponsored his application to the Eskom Power Plant Engineering Institute (EPPEI). EPPEI is a collaborative program between Eskom and local universities that aims to produce masters level research for the organisation. During his time with EPPEI, he completed his honors in mechanical engineering from the University of Pretoria, majoring in design optimisation and artificial intelligence. He also developed a novel method for clustering and structuring natural language operator logs. Realising the opportunity for practical deep learning products in South Africa, he left Eskom and co-founded Spookfish with Stephen. Robert holds a BSc. (Eng) Mechanical Engineering from the University of Cape Town.
Stephen also joined Eskom after completing his degree and spent two years in the Research, Testing and Development department. He was also accepted into the EPPEI program, and shared an office with Rob at the University of Pretoria. During his time at EPPEI, he completed his engineering honors, majoring in artificial intelligence and numerical systems. While his research was concerned with modelling pressure pipework for structural integrity, he spent all of his spare time upskilling himself in computer vision through an array of online resources. This self-taught computer vision knowledge now forms the technical backbone of Spookfish AutoAssess. Stephen holds a BSc. (Eng) Mechanical Engineering from the University of Cape Town.