Automation can unlock greater efficiency

Technology can help insurance companies control the huge amount of data they deal with every day.

Pamela Negosanti, global VP of insurance at Expert System, told delegates the vast majority of data insurers are dealing with is unstructured – possibly as much as 80 percent. She said a lot of knowledge-intensive documents are being used, along with a lot of cross-references. But a type of tech called natural language can help. It is being used to understand risks and make underwriting results more accurate and faster.

Negosanti was speaking as part of a panel at the Intelligent Automation event in London today (Tuesday May 21) during a session called ‘Exploit Natural Language Processing (NLP) and Natural Language Understanding (NLU) To Deliver Valuable Insights To Decision-Makers’, sponsored by Expert System.

James Breeze, digital artificial intelligence & analytics lead at AXA XL, was also on the panel, with David Clamp, managing director of Merlin Digital Consulting, moderating the discussion.

Expert Systems and AXA XL have been working on Cogito, an artificial intelligence technology that is able to think, understand and learn, and delegates were able to see a demonstration of it, with Negosanti claiming that it is the most accurate on the market.

According to Breeze, the main challenge is the number of documents insurers use every day, which means they have to rekey information. There will be automated submissions in the future along with shared services, he said, but stressed that it will still be important to understand what is in a document and to have an accurate understanding of the language within it.

Breeze added that AI can be the catalyst for documents to be cleaned up, so the data in them can be understood faster. While AXA XL is currently working on an insurance contract search system. He noted that it is important to have a well-thought through process that is sustainable and has processes connected in a logical manner.

Finally, the panel said that companies will be able to leave AI to deal with more mundane matters and concentrate on more important tasks to do with insurance risks. They said it is too early to tell if AI will completely replace underwriters in the future but said that in another five years the picture might be clearer.