#TECHVISION2020 Human-machine collaboration As these examples show, untethering creativity and catalyzing change begin These challenges are starting to disappear for text and spoken words at the level of human-machine collaboration. Insurers must think beyond a through advancements in NLP. Google’s BERT and Baidu’s ERNIE—which linear “command and response” relationship and engineer an interactive, are both open-source frameworks—enable AI systems to move from 22,23 exploratory and adaptable experience. This requires an innovative set of understanding just one word to understanding phrases in context. NLP capabilities that most enterprises aren’t actively building today. systems are increasingly able to understand the intent behind a human’s words, letting them better inform outcomes without constant intervention. Automation requires designing the skills to get a job done, but collaboration demands the ability to communicate and iterate with partners. Insurers Look at insurtech, Shift Technology, which has developed an AI solution to will need to explore and master the tools and advancements that enable detect claims anomalies and combat fraud. The solution analyzes structured humans and machines to better engage each other. Natural language and unstructured data from internal claims and contract data as well as processing (NLP), explainable AI and extended reality (XR) are examples of external sources such as meteorological data, satellite images or even tools that will all unlock new ways for humans and machines to interact. web data. The system uses NLP to extract structured data for use in fraud detection.Customers include the likes of Generali France, AXA and CNA 24, 25 Financial Services. Meaningful collaboration begins with communication, yet machines have historically struggled to understand human language. Machines typically are precise in their actions and operation, while language is anything but. Between slang, regional variations in dialect, sarcasm and single words that have multiple meanings,the challenges of language have complicated communication with machines. Technology Vision for Insurance 2020 | We, the Post-Digital People 25

Technology Vision for Insurance 2020 - Page 25 Technology Vision for Insurance 2020 Page 24 Page 26