Brokers wanting to take advantage of artificial intelligence and machine learning need large amounts of data and to hire the correct experts.
If a company wants to effectively build and deploy artificial intelligent and machine learning systems, it needs a large data set, Joe McKendrick wrote in The Data Paradox: Artificial Intelligence Needs Data; Data Needs AI, an article published June 27 in Forbes. The article is not about insurance, although it is in keeping with the industry’s discussion about AI.
“There may be ways to automate various pieces of data science roles, but the skills category that will still be essential is that of data engineer,” wrote McKendrick, a researcher who formerly served on the organizing committee for the Institute of Electrical and Electronics Engineers International Conference on Edge Computing. “There are many tasks required to source, manage and store data in which data scientists don’t necessarily want to get involved.”
AI involves technology that mimics “human cognition and activities” such as learning from experience and identifying patterns, Mark Breading, a partner with Strategy Meets Action, told Canadian Underwriter earlier.
The idea behind machine learning is to give a computer program access to volumes of data and let it learn about things such as the relationships between variables, University of Toronto Rotman School of Management professor John Hull said during the 2019 P&C Insurers’ risk management conference, held in Toronto.
Chatbots are one example of an AI application that brokers use. The AI part is natural language processing, in which the software interprets the meaning of a customer’s words without being specifically programmed to do so.
Insurers can use AI to save time and effort in handling documents, suggested Kamana Tripathi, vice president of global markets at TD Securities, during an earlier webinar. Also, claims departments can use AI to detect fraud.
McKendrick highlighted the need for a large data set to make AI and machine learning work.
“Data scientists and high-level data analysts will continue to be in demand, and are critical to helping enterprises design and test algorithms and data needed to predict trends, automate processes, understand customers, and engage with customers,” wrote McKendrick. “However, the amount of data flowing into and through enterprises is overwhelming, as are demands for new algorithms and capabilities — beyond what a data scientist can accomplish.”
“The most important role the most important first hire is a data engineer,” John Mosch, senior manager of analytics, business intelligence, and data science at Cisco, said as quoted by Forbes.