Data-Driven and Powered by AI

Businessmen studying graphs on an interactive screen in business meeting.

On the cutting edge of artificial intelligence (AI) and deep learning models, weather is the next frontier for technologists, who are analyzing patterns in a vast new stream of data. Along with a global pandemic, 2020 ushered in a year of extreme weather events across the U.S., ultimately causing an estimated $95 billion in property damage alone.

At Travelers, our leading capabilities in leveraging the power of AI and sophisticated modeling – combined with high-resolution aerial imagery and geospatial data – are delivering powerful benefits to customers when they need us most.

“We’re proud to be a data-driven company, and AI exponentially improves our capabilities,” said Travelers’ Executive Vice President and Chief Technology & Operations Officer, Mojgan Lefebvre, during a panel discussion at MIT about business uses of AI to improve customer experiences. “We also think about AI as a technique for our competitive advantage, especially when you combine high-quality proprietary data with third-party sources.”

Traditional analytics has always relied heavily on structured, well-organized data, whereas deep learning models now explore the expanding realm of large, unstructured data sets. These models can glean new insights from call center transcripts, identify property risks through high resolution aerial imagery or find patterns and correlations in notes taken when claims were processed.

“The way we are leveraging these deep learning models is connecting dots from large, nontraditional data sets,” added Mano Mannoochahr, Chief Data and Analytics Officer for Travelers. “Applying AI and deep learning models can unearth entirely new ways of learning from the past to reshape and plan for the future.”

Ever since the damaging spate of California wildfires in late 2019, for example, our technology and data science teams have leveraged deep learning to analyze thousands of aerial images of property damage – immediately identifying which properties are total losses. As a result, we can advance payments on most wildfire total-loss claims prior to having an in-person inspection, allowing our customers to begin the recovery process faster than ever before while keeping our employees safe.

More recently, we’ve used deep learning techniques in our Travelers Roof Shape Classification AI Model, currently available in some hurricane-prone Texas counties, to improve the speed and accuracy with which we can quote customers’ coverage needs specific to their roof type. The shape of a homeowner’s roof can make a big difference in how much damage a house sustains in high winds and thus impacts their insurance policy and the risk calculations involved in pricing.

Using the Roof Shape Classification AI Model, Travelers' technology teams rely on algorithms and aerial imagery to identify a roof’s shape – typically a time-consuming process for customers – with close to 90% accuracy. 

This advancement provides additional clarity and automatically completes portions of the customer application during the quote-and-issue process, removing a considerable barrier to entry for the customer.

Through the lens of AI and deep learning, our Chief Data and Analytics Officer sees an exploding ecosystem of emerging data sources and insights around us.

“This is helping us reimagine every part of our business,” Mannoochahr said. “Meaning we’re extending our advantage in risk expertise, improving productivity and efficiency of our business operations and greatly enhancing the experiences that our agents and customers have with Travelers.”

More Technology & Analytics

Future women in tech get coding experience at Travelers through Girls Who Code.

Collage of future women technologists attending a virtual conference.

Small groups within Claim work on innovative projects that yield meaningful business results.

Man and woman talking in an office.

Quoting and issuing new policies is easier and faster than ever, as Travelers BOP 2.0 transforms the way agents do business on Main Street.

Man looking at a computer screen.