Business AI Today: Hey People, You are Doing It All Wrong!

Jim Nelson, Rehn Professor of Business Analytics, Southern Illinois University, Carbondale

Way back in the 1990s, businesses knew that they needed more knowledgeable people to develop their information systems. The World Wide Web was rapidly expanding, e-commerce was the hot topic, and tech startups were everywhere. Since it all involved computers, businesses went to the most obvious source for more employees: University Computer Science Departments. The problem with that approach was coding is only about 10% of the entire systems development process. The coders they hired had no idea what the business was about, didn’t know how to collect requirements that fit the business, and couldn’t explain their systems to the business people. The business people couldn’t use their systems, and on and on it went. The code they got was amazing, fast, efficient, and a marvel to behold. And it was a horrible fit to the business problems it was supposed to fix. Businesses should have been looking at Business College Management Information Systems (MIS). They knew the business, they knew coding, and they knew how to make systems that could be used. It took the “dot-bomb” of the early 2000s for businesses to realize their mistake and concentrate on MIS graduates.

In 2015, Business Analytics was the new hot topic. Everyone wanted to do data-based decision making. There was enough data out there. Structured and unstructured data was overflowing. Organizations were saving every transaction and collecting massive amounts of data on their customers, employees, products, competition, and everything else they could think of. All this data, and what to do with it? Eager to get any advantage over the competition, organizations formed data science departments (either formally or informally) to analyze the data and to gain some insights into what the company was doing, what it did right and wrong, and what it should do in the future. The statisticians did amazing work turning out model after model. Management was impressed and the call went out for more data scientists.

Business colleges took notice of this new demand. They quickly created new courses, programs, and departments. But academic being academia, and not knowing the past and so being unable to learn from it, these business colleges renamed statistics classes as data science, and repackaged statistics majors (with a little database thrown in) as Data Science. They should have stopped there. But instead, they christened these new areas as Business Analytics because, after all, the programs were housed in business colleges and they had to have “business” in there somewhere. Businesses were thrilled and hired these fresh new data scientists/business analytics graduates.

Unfortunately, like the coders of old, these data scientists (I will not call them business analyticists) had no idea what the business was about. They were marvels with statistics turning out model after model that had fantastic statistical significance and beauty. But they had no idea what the problem was that they were solving, they had no idea what the data meant, and had no idea how to apply their models back into the business. They couldn’t even explain their models to the business people who requested them. But data science is only 10% of the entire business analytics process. No doubt it’s the “cool” 10%, but that ignores the hard work of truly understanding the business so the 10% could fit back into the business and actually fix problems and provide solutions. Organizations are going elsewhere or are growing their own true business analytics areas. Universities are abandoning their new data science areas as demand dries up. Today, only the true Business Analytics programs that integrate deep business knowledge with data science skills are surviving and thriving.

Today, the cycle repeats itself. Artificial intelligence (AI) is the hot topic. Everyone wants to develop, deploy, and dominate with the new Generative AI, AI robots, AI in every department and in every business function. Venture capitalists are throwing money at any business that can even spell “AI”. With AI specialists in such high demand, where do businesses go? Back to Computer Science, of course. They have the skills to build neural nets, deep learning systems, to train AI to do just about anything.

And once again universities and businesses are focusing on the cool, sexy, high demand fields of AI development. And once again it addresses only 10% of the problem. This is AI in business, not AI in isolation. There is still the very real issue of identifying business problems that AI can help solve. Otherwise, the business has a very cool AI system and has no idea what to do with it. And it sits on the shelf looking for something to do. In addition, the AI must be fed data. This data must be clean, curated, appropriate, unbiased, and constantly kept up to date. The AI solution must be monitored for correctness, efficiency, and effectiveness. It must be maintained. And on and on. All this is not glamorous. It’s very hard work. And it must be done with people who truly understand the business, its data, its functions, its goals, and its strategy. This is Business AI. This time around, let’s hope that we get it right.