Artificial Intelligence and Fresh Produce.

One of our major supermarkets has just announced a strategic partnership with Microsoft to develop it’s artificial intelligence capability (see computer world and financial review). If you are in the fresh produce industry this has some significant implications, which will be explained in this short article.

Their strategic aim can is to use artificial intelligence to become much smarter about how it manages it’s costs and value proposition to its clients. To understand how Coles will do that, you really need to understand what is meant by artificial intelligence.

Artificial intelligence is the ability of a computer program or a machine to think and learn. These programs don’t think the same way as we do, but in many respects they can learn a lot better than we can.

Thinking like a computer program
At the most basic level a computer program thinks by making a relationship between two objects. For example, if its yellow then its a banana and if it is not yellow it is something else. These if-then relationships when peeled back are just plain mathematics. This is how computers think. They think in terms of logic and mathematics.

Learning like a computer
To make the logic of mathematics work you need numbers. To make computer logic work you need data. So while humans learn through senses and experience, computers can only learn by accessing data. The logic can then kick in and solve problems. If the data is there, quite complex learning can take place with even simple if-then logic. For example if I have climate data, customer traffic data, and stock data, a computer can learn this – If its wet then I have less customers. If its wet the customers buy more soup. – therefore if its wet I might have less customers but need to stock more soup.

Why is it happening now?
As the cost of computing comes down, we can ask programs to do more complex mathematics without spending more money. The other cost that has come down is the cost of finding, collecting,and storing data. Thank the internet, your phone, and companies like google for that. With better mathematics and more data, computers are becoming better at learning. In fact computers can collect and analyse a lot more data than we can and much higher speeds.
However, we are not talking terminator. As the diagram below describes humans still have much more computational (thinking) power than computers. However computers have much more data analysing power. They can learn and find insights in massive data sources, that would simply be impossible for us to deal with. That is why artificial intelligence holds so much opportunity and the economics mean it is happening now and will only accelerate from here.

So turning back to Coles, why are they doing it?

The commercial returns you can achieve from a supermarket store is a function of many things. What is the weather like, what time of year is it, what is the traffic like, how much stock are you holding, what is the demographic of your neighbourhood, what time of day is it…and so on. These many data points will affect how much store traffic you get, how much stock you shift, and how many people you need to have at the cashier.

Why is this important for the fresh produce industry?

The reality today is that most supermarkets and independents maintain reasonable customer service levels by overstocking, and on occasion over correct and understock. As more stores use artificial intelligence to manage their inputs we will see significant changes in both the procurement volumes and the procurement methodologies that retail buyers use for managing fresh produce. This is not going to be isolated to Coles. This is what the future of procurement will be based on. To understand the drivers read this article from the progressive grocer

At the HiveXchange we believe the message that the market is sending is clear. As a supplier you must build diversity and resiliency into your supply chain. If the large retailers are going to use artificial intelligence effectively and get the commercial returns they expect, then the logic follows that they will want their supply chains to be more agile and responsive. They will expect to be able to move the right kind of fresh produce into their stores, in the right volume at the right time. No more, no less.

For downstream suppliers this will create new demands to find alternatives sources of supply when stock is not available and new markets when stock is in abundance. Our belief is that only digital markets can leverage the same computing infrastructure to learn quickly and find those relationships.