There’s a ton of hype around Artificial Intelligence (AI) these days. Companies like Microsoft and IBM are pouring dollars into advertising to demonstrate the power of AI, driving towards the ultimate goal of grabbing mindshare and business for tools, infrastructure, and products. There’s also considerable discussion about large scale AI ventures like self-driving vehicles and human-like robots.
With all this big picture discussion, it’s easy to neglect how analytics can be applied to relatively small scale, targeted scenarios. Pragmatic solutions that enable companies to use their data to make better decisions is where the real action needs to be today.
But the first step of using the data that you have is a challenging one. There’s a common heuristic in the data science world that 80% of the time is spent preparing the data – finding it, inventorying it, cleaning it, aligning it, and getting it into a accessible location.
Maturity models are a helpful tool for companies to evaluate a company’s readiness to leverage their data for decision making. In https://blogs.msdn.microsoft.com/stevengu/2018/11/01/the-ai-journey/, Steve Guggenheimer lays out one example of a progression from basic to advanced. I strongly support the first step in the journey: “BI before AI” – Business Insights (note, not Business Intelligence) before Artificial Intelligence.
The BI before AI step involves the pre-requisite of preparing your data. As noted above, this is a non-trivial step that takes some persistence. It also should be an iterative process – prepare, explore, learn… and repeat. Data preparation is the necessary step to get to the main goal – insights. Exploring the cleaned data through visualizations and basic statistics can provide ‘ah ha’ moments. Maybe the BI step is all the company needs today. Or perhaps the insights provide the inspiration for more advanced scenarios involving machine learning or other techniques.
BI before AI is an essential step towards becoming a data driven organization. Forget about robots and self-driving cars. Instead, leverage the insights from the data that you and/or your customers have available to drive better decisions today.
Picture details: Lida Late Afternoon, 12/9/2018, iPhone 7, F/1.8, 1/451 s, ISO-20