We should never underestimate the savviness, smartness and flexibility of the human. We can quiver as the front line of AI advances on many of our jobs, but in some areas us mere homo sapiens will hold our ground for a fair while yet. Strategic data analytics is one of those.

Strategic analytics is the discipline of using data to contribute to the key decisions of an organisation. This differs from tactical/prescriptive analytics which tends to be focused on specific use cases. No one doubts AI will be able to analyse data, build models and predict outcomes much better than any human analyst will do. But is it ready to take over the boardroom? Not quite yet.

So why are humans better equipped than AI in strategic analytics?

Repeatability : AI currently thrives in repeatable environments, in very sophisticated ways it can use the triggers that have caused past events to predict future ones. Most strategic analytical challenges are not that simplistic, not only having unique situations but usually unique underlying factors, nuances and contexts. This may be less so in online enterprises, but the greater the exposure to the real-world flesh and mortar, the more intrinsically unique each situation is. The human brain is remarkable in its ability to take such leaps, applying past insights and knowledge to radically different situations. As of yet, AI can't complete.

Complexity : Complexity in itself is no barrier, but much complexity is abstract. Success is balancing the knowns, the unknowns and the vague. Yes, CEOs are paid too much, but their value is their abilities to bring together these factors to make the right decisions by balancing purity, pragmatism, intuition and (ok) luck!

Data Accuracy and Assumptions : "It's what the data says" we often hear to justify decisions. But what does that data really say, after all, as Coase said, "if you torture data enough it will confess to anything". The role of a strategic analytics professional is not just to communicate what the data says, but increasingly what it does not say (i.e., call BS!). Data may be flawed, biased and have contextual nuances. AI is not yet in a position to grasp these well.

Implementation : AI can perform in astounding ways in virtual environments. But applying strategies to achieve goals is a different game, it is murky and uncertain. People are strange and unpredictable. But what they can do very well is respond to the virtual and physical, consider the strengths and limitations of humanity, and make it happen. Statistically pure and perfect is no match for what is practically good and works.

The Big Story : Strategic decisions are more often than not built into an underlying narrative of past, present and future. This pathway is rarely linear or even having any pattern at all. The story is not only key in making decisions, but key in building a culture to getting everyone behind making that story a reality. We have seen ChatGPT is good at making stories (but also making them up). As humans we are blessed with astounding vision.

Why is this important? For executives it remains vital to have an experienced human face to organisational data, analytics and insights in the board room. Such expertise that has full grasp of the strengths, weaknesses, nuances and contradictions in end-to-end data, around the lens of how the organisation works and the existing context. All executives should be equipped with these skills to some degree and asking these questions. But experts are essential.

This is not to say we should not be investing heavily in AI and related technologies. These have enormous power in contributing to all types of data analytics and making better decisions at all levels. This will continue at accelerating rates. But we are not yet in a place to fully outsource key decisions to our virtual buddies.

My advice to younger analysts is to continually hone your pragmatic and application skills and real-world experience. Our role is becoming like commercial pilots, less so much flying the plane but overseeing the journey, recognising the pitfalls and being there to take over when things go wrong. These skills are not only data-related, but the ability to present data, communicate insights and partner with decision makers. And perhaps most importantly, recognise pragmatic implementable solutions over simply the purest data solution. Building end-to-end skills and experience in different situations will serve you well.

​