Building effective Systems of Intelligence (SoIs) is a three-step process requiring business and design knowledge as well as data science and, of course, coding. In his latest Professional Alert, “Getting Started with Machine Learning Foundation of Systems of Intelligence”, Wikibon Big Data & Analytics Analyst George Gilbert identifies three steps to building effective SoIs, each requiring a combination of technical and business skills.

SoIs are built on machine intelligence, which means their potential capabilities are bounded by the limitations of the data available to them. Thus, the first step in the process of creating SoIs is to understand the available data, both from internal and external sources, in order to determine what is technically feasible. This is the province of data scientists.

The second step – prioritizing potential development projects – depends largely on their potential impact on the business and ROI. This step requires involvement of a business person. Gilbert warns that determining the answer to the question of which project to do first – something easy to accomplish or something with maximum impact on the business – is a more subtle process than it might seem.

The third step is designing the SoI. These are often highly customized and usually developed in-house. To maximize the impact of the finished SoI system, it needs to fit into an existing workflow of the business with minimal disruption. This puts design more in the realm of user experience and interface design informed by knowledge of the target business. SoIs tend to evolve over time as the quality and scope of the available data changes. Thus, unlike Systems of Record, which have a great deal of stability designed in and which impose a workflow on the organization, the most effective SoIs start as unobtrusively as possible and grow their influence over a workflow and its results over time.

 Copyright 2015 Wikibon, used with permission