AI and retAIl. The headline to this article is a joke. Or at least it is an attempt at one. There already is an A and an I in retail… lol.
I asked a friend: “Do you think this will be funny in Italian?”
He said: “Do you think it’s funny in English?”
One thing that definitely isn’t funny is artificial intelligence. It can’t write jokes. Here’s an example of a pun written by an AI: The greyhound stopped to get a hare cut. Oh dear.
Maybe that’s why many people fear that one day AI will kill us all – because it can’t get a laugh and it’s jealous of humanity’s sense of humour. Although if that does happen, it might take more than one day to kill us all.
One thing that AI is definitely great at though, is analysing data, spotting patterns and finding answers to problems. Actually, that’s three things. Anyway, what that means is before it decides that the Day of Reckoning has finally come, AI could actually be extremely useful in many applications.
And one of those applications will be in retail.
It is difficult to point out a positive consequence of the Covid-19 pandemic. It has, after all, been a complete menace, claiming more than a million lives, leaving millions of more people unwell, and unleashing economic disasters.
But for businesses in the ecommerce sector, there has been a colossal spike in demand. It has been truly unprecedented and in many cases extremely difficult to handle. Unable to get to physical stores, shoppers have gone online for pretty much everything and that’s put enormous pressure on fulfilment operations.
It’s important to acknowledge that the pandemic really is one of those once-in-a-lifetime events. But it has highlighted the need for flexibility and agility in businesses. The manner in which events unfolded was faster and more widespread than just about anyone could have foreseen.
This where AI steps in.
Imagine plugging all your datasets into an AI and asking it to create the most detailed and accurate demand forecast possible. It’s not exactly cutting-edge stuff, I know. You would probably get a thorough forecast produced in a fraction of the time. Which is good, right? But it’s not great.
Would such a thing help if there was global disruption even close to 2020 levels? Probably not. But where AI can make a real difference to forecasts through its pattern-spotting potential, and that could be a game-changer.
Now imagine all the links in your supply chain – your supplier’s suppliers, their suppliers, and on and on. If a fire or flood or revolution or epidemic causes one link in that chain to falter, the effects will be felt everywhere. Eventually. Those who spot the trouble first will be best placed to avoid it, switching to alternative suppliers before stock shortages hit the market.
The same approach could be used when tracking customer behaviour. Rather than just relying on past sales performance, more nuanced data from social media, meteorology, even news channels, can offer a richer insight into the mood of the public. Lots of people are watching cookery shows on TV. If there’s an explosion of interest in a particular item, ingredient or implement, pulling that information into real-time sales decisions can be invaluable.
Out on the road, trekking through the last mile, the same multi-dimensional data analysis can transform route planning, customer communications and much more. A major incident closes a road displacing huge volumes of traffic and causing congestion. But live traffic information could help automatically reroute deliveries, and notify customers of the delay – all in the blink of an eye.
None of this is beyond the capabilities of flesh and blood mortals, of course. But AI has a number of distinct advantages over us. One is the speed at which it can process information. Another is the extent to which it can find important patterns in the most unimportant-looking places and factor them into the decisions it is making and remaking constantly.
All of which is great. Except that it also means that when it finally decides to kill us all, it will be able to figure out where we’re all hiding and find us in no time at all.