Using automated machine learning for AI in insurance
Using dotData, an automated machine learning vendor, one of the largest insurance firms in Japan built out an AI platform that provides a personalized experience to customers.
Mitsui Sumitomo Insurance, one of the largest insurance firms in Japan, began the process of digital transformation several years ago. The company launched multiple projects, and continues to start new projects, to send it further into the digital age.
One of MSI's more ambitious undertakings is the MS1 Brain platform, an AI in insurance project to create a more personalized experience for customers.
AI in insurance
Released earlier this year, the MS1 Brain platform uses machine learning and predictive analytics, along with customer data, including contract details, accident information and lifestyle changes, to recommend products and services to customers based on their predicted needs.
"Our business model is B to B to C [business to business to consumer]. We provide our products through agencies," said Teruki Yokoyama, deputy manager of digital strategy in the department of digital business at MSI. "Until now, we have provided products to customers, both individuals and corporations mostly by leveraging experienced agents' intimate knowledge of client needs."
"By providing the needs analysis outcomes of each customer to the agency by MS1Brain, now even an inexperienced agency can make optimal proposals to customers with higher demands," he continued.
Mitsui Sumitomo Insurance used automated machine learning to build out a machine learning platform
Automated machine learning
MSI first connected with dotData in 2017, when MSI's CIO visited Silicon Valley for a technical survey, Yokoyama said.
At that time, dotData was just getting started, and it hadn't released a product. Still, MSI was intrigued by its automated machine learning platform, which claims to provide full-cycle machine learning automation. DotData competitors include DataRobot, H2O.ai and Auger.ai.
Automation of the data science process is the only way a company can truly deliver value from AI/ML investments and provide competitive differentiation by investing in predictive analytics.
Teruki YokoyamaDeputy manager of digital strategy, Mitsui Sumitomo Insurance
"When it comes to data analysis, model accuracy often gets the most attention; dotData, on the other hand, focuses on how quickly you can move from raw data to working models -- the AI-based feature engineering is what stood out," Yokoyama said.
MSI had to build a lot of intelligent models, said Ryohei Fujimaki, CEO and founder of dotData. But, the firm didn't have the data science team to build them.
DotData's platform was scalable and enabled MSI to automate the entire AI building process, from feature generation to model implementation, Yokoyama said.
"Everyone should embrace this approach," said Yokoyama of the automated machine learning approach.
"Automation of the data science process is the only way a company can truly deliver value from AI/ML investments and provide competitive differentiation by investing in predictive analytics," he said.
