"AI starts with data, and if the data is lousy, you're not going to make any great AI," said Bob Friday, co-founder and chief technology officer at AI-driven wireless network company Mist Systems, speaking on a panel about AI in IT. The data coming out of various firewalls, routers, load balancers and other devices that applications depend on varies based on vendors providing the equipment. But enterprises can make the most use of AI if the volumes of data extracted from the various elements within a network are formatted the same way regardless of origin, IT leaders said. Data standardization could be a precursor to large AI-based projects within IT infrastructure, and that conversation has already been going on within the IT community for some time. "Two years ago we were asking ourselves will we ever get to a standardized place … it sounds like that's still not settled," said Neal Secher, senior vice president and head of networks and data center modernization at State Street Corp., at the panel event.