Managers manage. And managers decide. Even as AI experts are jumping in with predictions of when artificial intelligence will have completely replaced humans, in management circles human decision-making still holds sway. But what was once customary is increasingly becoming problematic, because it's ever harder to gain a full overview of business situations.
"The human brain quickly reaches its limits in complex planning and decision-making tasks. This is particularly true when it comes to tactical management of operational processes, such as manufacturing, transport logistics and supply chain planning," says Adrian Weiler, CEO of the Aachen-based company Inform. "Thousands of activities have to be arranged in the best possible order with hundreds of resources allocated to them. There are uncountable billions of options for such arrangements. A person can never have a full overview in this kind of decision-making space, so they'll limit themselves to the plans that appear plausible." And yet what appears plausible from a limited appreciation of the facts may not be anywhere near the right choice.
Another land mine comes into play as well, according to Weiler: The world keeps turning. "Even if managers find a good plan, they need to constantly be adapting it, because ever more disruptive factors impact these plans in our globalized and increasingly networked industry," adds the AI expert. Managers are very naturally challenged in this situation. And that has economic consequences.
Take a mechanical engineering firm. If it doesn't manage its orders optimally, important technicians might not be available for assembly. Or parts might be stored in warehouses, tying up capital without being used. A logistics expert who doesn't optimize their routes risks having their truck fleets travel unnecessary extra distances.
"Things are particularly problematic when daily operations aren't running smoothly due to an unplanned disruption," warns Weiler. An urgent order comes in or a machine breaks down unexpectedly. The short-term planning revisions that are then needed, involving all impacted business resources, become a feat that even the best manager can't solve manually. If the part leaves too late, the company has extra costs to contend with. And far worse: Their reputation is at stake. Damages to image can hardly be reduced to a number.
But there's no need for things to get so bad. "Mathematical optimization algorithms offer a complete overview of all action choices for a decision, and present a good alternative to analog management decision-making," says Weiler in recommending the use of artificial intelligence. And not just for some sectors. On the contrary: Complex decisions are increasingly needed in every business, from carmakers to container terminals. Artificial intelligence is a viable option anywhere where multiple parameters and fast change make decision-making particularly complex, in production, for transportation routes, warehouse management or sales planning.
"Instead of making the decision themselves, in the future managers will be free to experiment with the various resource figures. Because using if-then analyses, algorithms can calculate in seconds how one employee more or one machine less impacts the overall result," says Weiler.
These "if-then analyses" are also a valuable support in creating offers. For example, if an event manager wants to see whether 100 employees with very specific qualifications are available at a given time for a large conference, algorithms help with a short-term if-then analysis to ensure a fast and reliable offer.
If algorithms work so well, why do managers have such a hard time leaving their decision-making to a machine? "To accept optimization, you have to leave plausibility behind," says Weiler. "That's hard for managers to do. They don't have the necessary trust." Whereas in their personal lives this is no problem. These days people put their travel destination into their GPS or Google Maps without hesitation – and trust that the best possible route will be given them. While this trust has become commonplace in our daily lives, business executives still hold tight to their familiar, analog decision-making methods. And yet for the goal of "more productivity" or "faster processes," there are often billions of possible approaches, and mathematical optimization algorithms can automatically work out the most efficient of them.
"In the future, healthy common sense will be called on more for strategic decision-making," Weiler believes. "Optimization algorithms can find the best decision, but only within a grid that the manager provides." Goal-setting remains as always a human task. Management should concern itself above all with the question of where they want to go – and continue to refine this goal strategically based on ever more efficient processes.
So it's not about pushing people out, but about properly distributing decision-making power to wherever each is better suited to support the best performance. To Weiler it's clear: "Managers decide where to go, and algorithms choose the most efficient way to get there."
Making decisions at lightning speed does no good at all if the business structure still runs analog. Dynamic decision implementation is key. "Only when both come together – the 'best option' identified through optimization, and structural agility – can the company adapt to any new market situation in the shortest possible time," says Weiler. Still, he advises against a sweeping transformation: "Managers should first try out artificial intelligence for certain planning processes, and gain experience. If the strategy proves itself – and it will – then agile optimization can be slowly expanded and developed from the bottom up." When decision-making processes result in more productivity and customer satisfaction, then new doors will automatically open within the company, and new application possibilities will appear. "To remain competitive over the long term, it's important to take the first steps now. It can be a very small one to start," says Weiler.