What You Need to Know About Adopting Big Data, AI, and Machine Learning

Man and robot shaking hands

Most businesses are now very familiar with technologies like big data analytics, artificial intelligence (AI) and machine learning (ML). Executives have poured over case studies and performance statistics to better understand how these technologies can improve their products and services, as well as improve efficiencies and reduce costs associated with their own internal processes and operations.

It is true that there are companies using these technologies, but not as many as you might think based on the benefits that these technologies can bring. A recent survey by RELX asked 1,000 senior executives with companies across the United States for their thoughts and plans regarding AI/ML and data analytics. Eighty-eight percent of the executives responding to the survey agreed AI, ML and data analytics would help their businesses be more competitive. However, when asked how many had actually implemented these technologies, only 56 percent said they had. As for future plans for increasing the use of AI, ML and data analytics, only 18 percent said they planned to increase spending on these technologies moving forward.

Given the benefits, why isn’t everyone using data analytics, AI or ML? I believe this gap between the understanding of the benefits of these technologies and their adoption rates results from a combination of four issues and misperceptions. All four must be addressed by companies looking to adopt these technologies BEFORE they start implementing them.

Skills Gap

AI, ML and big data are relatively new and require skillsets that are uncommon in most IT departments. For example, proper analysis of big data requires a data scientist, which is someone with good understanding of math and data analysis techniques, business knowledge and some programming skills. How many IT departments do you know that have a data scientist on staff? In some cases, the lack of a data scientist can be addressed by a multidisciplinary team, assuming that the organization has individuals with the right expertise in different areas but, still, how many organizations can count these types of individuals among their employees and manage to drive these types of projects with multidisciplinary teams?

Domain Expertise

While a computer is certainly faster than a human is when it comes to making calculations, they are not necessarily capable of making better decisions than a human makes. To illustrate, imagine two baseball pitchers: an inexperienced rookie in top physical condition and a cagey veteran with years of experience. While the rookie can throw faster than the veteran, he doesn’t have the expertise needed to know which batters in the league have trouble hitting a curveball. That kind of insight only comes through having faced those batter multiple times, not from athletic ability. Businesses are the same, and no matter how fast or advanced an AI algorithm is, without an experienced human providing it with the right data, the value of the results the algorithm delivers are greatly reduced and, more important, the interpretation of these results can make a whole lot of difference.

Fear, Uncertainty and Doubt

I’ve previously written about the fears people have about AI and why I think they’re misplaced. In short, while AI and data analytics run on computers capable of much faster performance than humans, they lack the domain expertise I describe above to truly replace a human expert. Critical thinking and expertise require years of experience and a thought process that computers simply can’t perform as well as humans. In particular, intention, innovative thinking, and the ability to use holistic approaches potentially ignoring short term local benefits in lieu of larger strategic rewards are all characteristics of humans that computers have yet to replicate. So employees worried about a computer taking their jobs away should be reassured that Technologies like AI and big data won’t replace them, they’ll just make their jobs easier (at least, for now).

Cost

There’s a perception that the more cutting-edge a technology is, the more expensive it is to implement. So when a business development team starts talking about exploring the benefits of AI and data analytics, IT executives immediately start to calculate the associated costs: new hardware purchases, expensive per-seat software licenses and the pricey systems integration consultants needed to get these technologies up and running. The good news is that many companies are developing AI and data analysis solutions using open source software. Such software is free to anyone that wants to use it. Better yet, open source software gives companies access to a global workforce of software developers who can help them customize their AI and big data technologies and trouble shoot problems as they arise and these improvement can be fed back into the open source code repositories for the benefit of the entire user community. This keeps businesses from relying on the whims of a proprietary software vendor to release the bug fixes or new features a company needs to get the most from their AI, ML and big data technology implementations and avoids the dreaded vendor lock-in situation.

This article originally appeared on MoneyInc.com on March 13, 2019.