As AI moves out of research labs and into the real world of commercial applications, we will increasingly see the rise of AI products. Whether it be detecting fraud in financial transactions or optimizing supply chains, while you won’t need a detailed knowledge of machine learning models to take advantage of the next generation of AI tools, you may well require an appreciation for confidence intervals and a new approach to making decisions. On a trip to Toronto, I caught up with Karthik Ramakrishnan, Head of Industry Solutions & Advisory at Element AI. We spoke about the near-term challenges of embedding AI decision-making in organizations, and why just as important as getting algorithmic products to work with people, will be getting the products to work with each other to make complex, synthesized decisions across the company of the future.