The goal is to develop tools that ultimately indirectly are fed off and developed by access to that data. I use the expression “supply chain”. At the beginning, you have fundamental techniques, algorithms and basic hunches, of how to build the product or provide a solution.
If you want specific examples, we can think of development of tools for the environment. We want to develop a tool that can better predict local weather patterns, perhaps tornadoes in the Ottawa area. That's a great topic of note recently. In agriculture, you could want to develop tools that improve crop yields or get smarter about environmental impacts. From there you're going to look at diverse data sources to develop tools. Maybe the deciding factor is when to sow seeds in a field. That's what the algorithm is going to look for. It will come from diverse sources of information to get to that answer.
That's where AI is truly powerful because it can combine and apply this wide-ranging data and then come to a more digestible answer in a way that human minds cannot necessarily do.