Many Supply Chain Executives are nowadays questioning what AI (Artificial Intelligence) in Supply Chain is. It is still the “stuff of the future” for some organizations and to some history’s biggest paradigm shift. Yet, we are still only at the beginning of the AI (r)evolution. In this blog post, we will look into what is AI in Supply Chain.
What is AI (Artificial Intelligence) in the Supply Chain? To some Supply Chain Executives it is still a mystery. How can this technology provide a competitive advantage to the Supply Chain? To learn more about how you can gain a competitive advantage by using AI in the Supply Chain you can read this blog post. In the following, you can learn more about the basics about AI and through an example learn about the logic behind.
The basics of AI in the Supply Chain
In its simplest form, artificial intelligence is when a computer or computer-controlled device runs functions and performs tasks that would normally need human assistance. Depending on its use, this may require a computer to understand speech, have a visual perception, or be able to make decisions that are independent of human intelligence (aka manual intervention).
Many AI machines can also learn on their own and make decisions based on past events. Or, they can interpret massive amounts of data from a company’s supply chain history to predict future outcomes with incredible accuracy.
Terminologies used within AI:
- Artificial Intelligence: A program that can sense, reason, act and adapt
- Machine learning: Algorithms whose performance improve as they are exposed to more data over time
- Deep learning: Subset of machine learning in which multilayered neural networks learn from vast amounts of data
The following three trends have put AI within everyone’s reach.
- Costs for data storage and processing has gone down
- Data availability has increased, not only inside the company but also outside
- Advanced mathematics has enabled complex data modeling
Though the concept of AI is half a century old, interest in the technology has snowballed over the past decade. This has been driven particularly by increased data availability, a lower cost of data processing, and new advanced mathematics, allowing for a significantly lower cost of predictions and analytical capacity.
In Optilon we are seeing particularly fast growth in three areas of AI:
- Text and numbers: Today, a wealth of information can be found in text and number formats.
- Speech: AI can also be used for speech recognition, which translates spoken words into text.
- Images: One of the most common uses of AI is image recognition. AI can be used to categorize, edit, and parse this image data.
Explaining AI using examples
In order to take out some of the mystery which is surrounding AI let us take an example from outside the Supply Chain.
The Real Estate Agent
Imagine a real estate agent who sells apartments. Every time a new seller signs up, the first task is to suggest how much their apartment can be sold for. The real estate agent is busy, and it is time-consuming to inspect all those apartments, so he decides to automate the task using artificial intelligence. The formula behind the intelligence sounds simple. Price is equal to the number of square meters times 20,000. Therefore, an apartment of 100 square meters is always priced at two million. After a short while, the real estate agent discovers the challenges of using this method.
Obviously, there are other things than just land, which affects the price. As an example, ground floor apartments are typically not nearly as expensive as apartments with views. Therefore, the next generation of intelligence puts a sum for each floor apartment located above ground level.
After looking at data for actual sales prices, the formula is refined a bit, because it is mainly on the lower floors that the distance from street level makes a difference. In other words, there is a greater difference in price between the living room and the first floor than between the sixth and seventh floors. In the real estate agent’s formula, it is fixed by taking the square root of the floor number and multiplying by DKK 100,000. Then an apartment on the first floor will be priced DKK 100,000 higher than the ground floor, while one on the ninth floor will be DKK 300,000 higher.
The following generations of formula incorporate further conditions: Whether there is a lift in the property, how attractive the area the apartment is in, whether the apartment is in good condition, how busy the road is, and whether one can see the sea.
Along the way, the formula has become somewhat longer and somewhat more complicated. It has been necessary to use both square roots and logarithms, but it can still be shown on an A4 sheet of paper if you use small print. One can question whether it is reasonable to call the formulas in the real estate example and software in self-propelled cars “intelligent”. As we have seen, in both cases, these are just formulas. Very long formulas, but still only formulas.
They are not, in principle, different from the functional machine whose intakes consisted of the formula y = 2x + 1. Artificial intelligence is, therefore, in no way intelligent. Absolutely not. The name probably describes researchers’ wet dreams about what technique one day they will be able to develop themselves. Not what it actually is.
So now, we hope we have taken out all the mystery, and you should be ready to look at how you can incorporate it into your own Supply Chain.
