Companies that understand how to apply machine intelligence will scale and win their respective markets over the next decade. That said, delivering on this promise is much harder than most executives realize. Without large amounts of labeled training data, solving most AI problems isn’t possible. The talent and leadership to bridge the worlds of product design, machine learning research, and user experience are scarce. Many organizations will tackle the wrong problems and fail to ship successful AI products that matter to their customers. Pete Skomoroch explains how to navigate these challenges and build a business where every product interaction benefits from your investment in machine intelligence. This talk was presented at the 2019 Strata Data Conference in London. Topics include: Who defines the data vision and roadmap in your organization? Who is accountable for building and expanding your competitive moat? Investing in foundational data infrastructure, training, logging, and tools Fostering executive support for exploration and innovation, including user-facing data product and algorithm development How to evaluate new machine intelligence projects and develop a portfolio that delivers How AI product management differs from traditional product management How to bridge the worlds of design and machine learning to get to product-market fit Defining a framework for trading off investments in data quality, machine learning relevance, and other business objectives