DECEMBER 2019CIOAPPLICATIONSEUROPE.COM9Such capabilities deliver decision support systems that operate with mathematical precision, something that was impossible a few years ago.Ideally, these types of analytical function complement each other and operate on shared data infrastructure. Each function, however, presents its own unique challenges such as utilizing data in different ways, requiring different types of talent, and leveraging different methods or analytical tools.Crucial for all functions is a need for strong collaboration between relevant business stakeholders to enable knowledge transfer in both directions. It is the imperative data scientists who can develop an in-depth understanding of relevant business areas. While at the same time, business stakeholders must discover what AI systems can and cannot solve.A healthy data landscape will benefit all three types of analytical functions but in different ways. While data serves as a repository for information and insight generation, as in the case of (I) and (II), it directly drives the performance of innovative digital solutions and AI-systems in (III). The more the data available and higher the quality of that data, the more accurate will be the predictive power of a recommender system that utilises it. Internal development of such smart solutions is becoming an increasing necessity to secure business opportunities, boost competitiveness, and build IP. New legislative frameworks and regulations can render opportunities to adopt competitive positions as well in the future.Within the blink of an eye, AI-solutions can utilise data that would have taken human a decade to merely read through or understand. In such cases where data is readily available, it is important to understand through practical experience what AI systems can and cannot do. As Andrew Ng said in 2016: "Anything that a typical human can do with one second of thought can probably be automated with AI now or soon."Pre-existing organisational structures, processes, talent, or other factors, including strategic consideration, might render more feasible approaches to corporate analytics than other methods. However, each organisation will have different demands and ultimately will have to find the solution that best satisfies their specific requirements.
< Page 8 | Page 10 >