The Machine and I
With all the hype and talk about Machine Learning and Artificial Intelligence (AI), it is easy for one to get lost in the ocean of information. Before you go running head long into a decision that is going to drain capital from your business and more importantly waste your most valuable asset, time, of your people. It is a good idea to step back from the hype and ask one question.
As businesses in our markets we are looking for anything to give us a sustainable competitive advantage. We are looking for what the U.S. military calls a “force multiplier”. This is some process or method or tool that would allow us to do more with less, reduce risk, and optimize outcomes. This was seen with electricity, the steam engine, the motorization of the work force, the Internet and connected computers. The case can be made for what the current hype cycle is calling machine learning, deep learning or artificial intelligence (AI).
But so what? With any technology, in most cases, it is best to understand what is being asked? What problem is trying to be solved? What process can be improved? Where can cost be cut to gain that advantage? And ultimately, does it make sense to use machine learning to do help solve these problems, answer these questions?
Let’s be clear what machine learning is and more importantly what it is not. At the core machine learning is the process in which humans are feeding data, text, numbers, images and sounds, into a larger networked computer to apply statistical models to find insight in that information. Insights or outcomes may be
- If we reduce cost in this market by three percent there is a probability of increasing profits by ten percent.
- If we run the factory during these hours and do preventive maintenance on the machines there is a five percent chance it will prolong the life of the machine by three years.
- If we increase our adverting and social media budget by fifteen percent we might be able to gain market share by thirty percent.
The type of business questions that actually help the firm. Another question that needs to consider is how do I get there from here? What is the cost in both currency and time? Much like many business, unless you’re a large technology firm like Google, Facebook, Amazon, or Microsoft, your company probably doesn’t have a budget or team to do machine learning. There probably aren’t data scientist on staff. And your IT department is already overwhelmed with much of the day to day operations that they are trying to stay ahead of for your business. So where do you start?