AI and Analytics at Scale

Chapter 3. AI and Analytics Together

AI is at its best when it does not stand alone.

Because AI is a new approach for many organizations, some people mistakenly assume it requires entirely specialized resources and skills, something separate from mainstream business and organizational processes. Furthermore, people who are working with data infrastructure that imposes unnecessary limitations may also mistakenly assume that AI must be built and deployed as a system separate from analytics applications. These ideas are not only incorrect; they also lead to unfortunate trade-offs in cost, performance, and flexibility. Some aspects of these trade-offs are specifically harmful, and all are avoidable.

This chapter is not a step-by-step guide to building AI systems. Instead, it addresses some key approaches that can make AI successful in practical business terms. To do this, the chapter explains not only how you can put AI and analytics together but also why you should.

Why AI and Analytics Together?

Putting AI and analytics projects together on the same system is good basic business practice. Running these applications on the same system makes sense because it is more cost-effective. Hardware and software resources are not unnecessarily duplicated. Furthermore, running fewer separate systems means less burden on IT and reduces the risks imposed by having many separate security systems. Conversely, separating these systems makes the logistics of handling the data shared between these ...