Spare parts planning: merging artificial and human intelligence

In the era of emerging technologies such as AI and IoT, more and more decisions in supply chains and logistics are being automated and systems, machines and devices connected to the web. However, the human brain is still crucial in taking proper planning decisions. In fact, I foresee collaborations between artificial and human intelligence.

AI: learning from past experience

With great show cases such as self-driving cars, face recognition and search algorithms, artificial intelligence is becoming more and more mainstream in society. However, there still is some criticism of these concepts. As AI is really just a few lines of code, it offers no creativity, does not comply to ethical standards and it can only learn from events that happened somewhere in the past. This also applies to for planning decisions in spare parts management. Based on tons of meta data, an AI model can give you a good planning decision, but what if we are dealing with a part that has its specific story and stands on its own? Is the learning curve that steep?

Human planners can judge

In situations where patterns in data can be found easily AI outcomes are often reliable. Unfortunately in spare parts management these patterns, e.g. in failure data, engineering data or supply data, are not so stable. In other words, finding a one-off event is quite easy. Think of the decision to stock or not stock an extremely slow moving part for a specialised asset in the mining industry. This is where human planners come in. They have vast business knowledge and are able to judge the consequences even when particular situations are edge cases. These planners can also assess when there is more information or consultation needed to take the best planning decision in that situation.

Bounded rationality makes it tricky

Even when we deal with seasoned experts in planning or engineering departments, there is still the concept of “bounded rationality” to keep in mind. In fact this concept tells us that people almost always take decisions with the “limited cognitive capability of the mind, and the time available to make the decision”. In spare parts management it is crucial to weigh different performance indicators to qualify a planning decision as “good” or even “optimal”. Think of elements such as availability, stock holding and operational costs. When a planner has a strong Finance focus, the planning decisions are likely to do more justice to the cost element than to availability. This often happens unconsciously.

Let’s join human and artificial intelligence

My point is that we should combine forces. Filter the value add of both artificial and human intellect. In practice that means we should use AI assistance. These assistants give you suggestions for planning decisions that a human planner can flavour if he or she thinks an update is needed. A nice side effect is that with every update or approval of a suggestion, the AI model is trained and thus improves future suggestions.

Lanza is a web-based spare parts planning solution that is aiming for the best possible planning decisions. Embedding AI assists into our software is part of development strategy of Lanza.

Most likely you are involved with AI technology as well. What do you believe in: AI automation or AI assist? Let me know!

Also check

Meer weten?

Stijn Wouters
Product Manager