Spinrite

Who is going to argue that artificial intelligence (AI) is, by far, the hottest topic in the tech and data analytics industries and maybe one of the overall trending topics worldwide?

Some people are more into tech and technological changes than others. If there were a scale to measure how much someone is conservative or enthusiastic about technological changes, I would be closer to the middle towards the enthusiastic side. Although I am not the most avid for technological changes, I can say that AI has already changed my professional and personal lives positively.

All this to say that what is next could sound like I am an anti-AI type of guy, which is not the case. I am just a data analytics professional who has been in the industry for almost 20 years. What is going to be shared is just the result of observation over all those years. The following few paragraphs are a collection of thoughts, the fruit of years working for multiple companies in multiple industries, doing consulting, being a professor in the field, and, of course, networking and exchanging experiences with colleagues around the world.

But if AI is that much of a trending topic and has started changing, in a good way, people's and professionals' lives, where are the contradictions mentioned in this article's title?

They reside in the fact that while companies are discussing how to leverage AI, some old historical data struggles did not go anywhere, and these are being traded-off for AI, which is where decision makers have more willingness to invest – here thought, but traditional business intelligence projects, focused on building data warehouses, and reporting/dashboarding platforms, have cost very much for organizations and, at many times, failed. There might be erroneous conclusions that this type of project simply does not work, but some of the mentioned struggles would only be solved by them. It is a matter of ‘and’, not ‘or’.

Well, let’s finally dive into these historical struggles and see important aspects of data analytics which should not be obfuscated by AI.

"Managers: AI is just the tip of the iceberg."

This starts with data accessibility. By that, we can understand having actual access to usable data, which in other words, means that data is clean enough that can be used for building visualizations, running machine learning models, data analysis, etc. Recent researches say that between 40 to 80 percent of a data analytics professional time is still spent either collecting or cleaning data. This includes different professions like data analysts, business analysts, data scientists, and others. This is a huge efficiency problem. Before thinking of algorithms, models, problem-solving, insights, and recommendations, these professionals – who can be very expensive – spend 40-80 percent of their time collecting and cleaning data.

Data accessibility is just the foundation stone of what is commonly referred to as a data analytics maturity framework. Simply by searching for this term online, it is possible to find multiple sources where a horizontal axis shows different levels of data analytics maturity. It starts with descriptive (what happened?), diagnostic (Why did it happen?), predictive (What will happen?), prescriptive (How to make it happen?), and cognitive (AI). On the vertical axis is business values. An exponential line chart links both axes and shows how much value is generated for a business when climbing the maturity level. Well, when most of the attention is on AI, who is going to ensure daily business questions like ‘What happened?’, ‘Why did it happen?’ and so on are answered? This still is a struggle at different organizations at different levels and forms.

The data analytics journey through the maturity levels can be bumpy depending on the data analytics culture of an organization. It is hard to determine who came first, a strong culture that built a mature data analytics platform, or vice-versa. But the best question is: what is a well-established data analytics culture? Business stakeholders are usually the most empowered and knowledgeable people to make decisions in their own business, which is kind of obvious. Data analytics is not around to replace that expertise. But to empower these business teams with access to actionable insights derived from data through visualizations, reports, dashboards, research, and analyses, which serve as resources for making decisions. When most business decisions are made like that, then we can say we are in front of a well-established data-driven culture. This usually requires constant effort from business professionals at all levels.

The role of the ‘middle-person’: data analytics professionals. As mentioned above, there are many types of roles in data analytics. If we were to combine all of them in one data unicorn, it would be a professional who can connect and translate problems, questions, and solutions, both ways, end-to-end, from technical to business. This is a very valuable resource, but many organizations have this ignored, and their data analytics strategy ran by a technical or a business department. It can work, but not ideal.

I will conclude this analysis by directing messages (provocations, actually) to different stakeholder groups to push them out of their comfort zones. If you get to this point, I invite you to reflect on them.

IT/Tech professionals: Big data and analytics platforms and solutions are fascinating and usually drive most of your interest and attention. I invite you to spend just a little more time talking about business problems with business teams.

Data analytics professionals: we are fascinated by solutions, models, and platforms as well. I invite you to spend more than just a little time talking about business problems with business teams.

Business teams: you usually know what numbers, charts, or reports you need. I invite you to bring up the questions you are trying to solve first. Then we’ll discuss numbers and reports.

Managers: AI is just the tip of the iceberg.