We tend to be afraid of things we don’t understand. With the proliferation of AI, people commonly lament that “the robots are taking over!” Luckily for us, we aren’t looking at a Terminator or Wall-E situation, but somewhere in between.
The value in machine learning isn’t necessarily that you can simply teach machines things and they’ll do them. One of the real values of machine learning is the ability to grow into a scenario where the machine is your eyes and ears that can enable organizations to better identify underlying events and concepts that are occurring. The sheer fact that you’re using a machine learning pattern or capability means you’re becoming more aware of how your data works and how your business works. By focusing in on the information in front of you and taking a data-driven approach to business, organizations are able to make valuable business decisions backed by data.
Organizations oftentimes can get off track by trying to accomplish greatness in one fell swoop, which is not the best approach. Of course, what greatness looks like exactly will depend on your company’s goals and objectives, but most organizations will point to some combination of being a pioneer or having an impact on society in a positive way.
Reaching that level won’t all be sunshine and rainbows. However, even an incremental gain is a big deal. Start with core fundamental patterns and practices. Are there repetitive tasks the team is performing manually that can be automated? That will free up more time to work on additional improvements.
You don’t need to have Watson or HAL running your systems. Simple insights can lead to monumental improvements in process. But how do I know what’s interesting? How do I compare data? What should I be looking for?
From an operational perspective, machine learning can help answer those questions. These systems produce tons of data, so it’s best to start simple. Focus on specific primary level events that form around a limited set of discrete variables. Your IT team should be able to identify a scenario and then react based on those insights.
For example, you may not know your marketing team runs a promotion every Tuesday before Christmas, but the data shows a spike then. This can help businesses be better prepared for this instance and anticipate issues that may be coming in.
Machine learning can make this all easier and it’s increased the breadth of automated tools to help you see things you don’t see naturally. It’s making the data more accessible for all teams, and that can help drive efficiency across the board.
While this democratization of data is exciting, it’s not a silver bullet to success. We have to be looking for opportunities to use machine learning. Think of how Netflix suggests new shows for you to watch, or how Amazon recommends products based on previous purchases. Those are great examples of companies finding areas to use machine learning to further serve customers.
This process is successful because you’re engaged in the exercise of training the system by actively evaluating the performance of the process within a feedback loop. That feedback loop drives further detail into the learning algorithms about what variables are interesting for additional analysis.
In his book Good to Great, Jim Collins takes a look at why certain companies thrive while others merely exist. One quote that resonates with me is when the “challenge becomes not opportunity creation, but opportunity selection.”
Opportunities to improve exist at every organization. We’ll find places that are a great application of use, and others that won’t be so helpful. The key is focusing on how we innovate forward.
Datapipe has joined forces with Rackspace to create the world’s leader in multi-cloud managed services. Learn more about the acquisition here.