The set of people who are frustrated every day by badly designed information technology is very large. So is the set of people whose dollars pay for the badly designed technology. A conservative estimate ranges in the billions for the cost of large-scale information systems that end up collecting dust because they're not properly human-centered. Yes, billions and still counting-that's the scary part. Within this large set of frustrated customers (see the sidebar "When Systems Development Neglects Human Considerations") is a subset whose job it is to do something about this situation. That subset includes policymakers, program managers, and systems engineers. It also includes a sub-subset comprising cognitive systems engineers, ethnographers, and many others who, in one vernacular or another, advocate human-centered computing. We must show that intelligent technologies-those designed to interact with humans or play a role in the cognitive work conducted in sociotechnical work systems-are usable, useful, and understandable.