Being able to derive insights from data has been a differentiator for a number of the most successful startups (Canva, Netflix, and even Amazon), but when starting a company, many things can seem more important than cataloguing and retaining data. In this session, we share guidelines on how to plan ahead for AI, including what kinds of problems lend themselves to being addressed with machine learning, what data to prioritise and what to throw out, and how to best structure and cost-effectively store the most valuable information that may provide long-term benefits.