Have you noticed? We’re on the cusp of a new age—one of driverless cars and endless automation. It may still be years (even decades) away, but machine learning and other forms of artificial intelligence have already become a business necessity for many organizations. As IT managers, you need to be prepared.
Machine learning put to work
On the most basic level, machine learning is all about computers learning from the data they’re given and making predictions, or decisions, based on that. It’s a technology we all interact with—it’s what allows for the personalization of your Google search results, Facebook News Feed, and Amazon product recommendations.
That’s just the tip of the iceberg, though. This sort of technology is making an even bigger impact behind the scenes. More than 40 percent of executives at large companies say their organizations are already using machine learning in sales and marketing, according to a survey conducted by the Accenture Institute for High Performance. Expect that number to grow—76 percent of respondents to that survey said they think the technology will help them achieve higher sales growth. In addition, it’s being used to help veterinarians diagnose patients. It’s helping banks recognize customers when they call, and it’s being used to reduce vehicle accidents—and the associated costs.
Much of this adoption is being driven by the rise of the Internet of Things (IoT). All those connected devices are generating massive amounts of data—too much for even the largest team to go through—and much of that data is unstructured. Artificial intelligence allows you to make sense of that unstructured data.
A matter of quality
Still, implementing machine learning is a challenge. Companies need to know what business problems they’re hoping to solve, the causes of those problem, and what data they have around those causes. Without the right data, even the best AI can’t do much. That data also has to be good quality, or “clean.” Problems with data will lead to problems with results. Implementing this tech isn’t a fast process—these algorithms need to be trained, and that takes time.
Depending on the business problem the company is looking at, an off-the-shelf solution might be the answer. For others, there are a wide variety of APIs that can bring this sort of power to another program. IT managers will have to help their organizations navigate these challenges.
IT managers will also have to think about the infrastructure they’re using—AI systems tend to require a lot of computing power and the massive data sets that feed into them will have to be stored somewhere. For some organizations, the solution will be in the cloud. That means IT managers will have to deal with the same issues that come up when using any cloud-based service. With much of the data feeding into these systems coming from IoT devices, data security will also be a major concern.
While machine learning may replace human workers in the future, that’s still (mostly) the stuff of science fiction. The current reality is that this technology is more of an assistant, helping people make decisions based on data sets that are just too big for them to deal with. That said, preparedness is a virtue—especially when it comes to new technologies.