Why AI is good for retail and good for jobs

The media is telling everyone that AI is going to create human redundancy on a scale not yet seen, because so much of what humans do now, can already be done better, faster, cheaper by technology, specifically Artificial Intelligence and its afterthought, machine learning.

That’s true, but the media then stops there because its human interest remit limits it to only caring about the impact on humans. AI is a timely reminder that it’s called the human race for a reason. It is a race, and one in which the winners are getting faster and the losers are often not even getting off the blocks. That’s just a fact and don’t kid yourself that capitalism and its enforcers, one of which the most faithful is technology, care about the expanding gap between the two.

When Radio 4 featured AI last week, they tried to get some retailers to talk, but you can hardly blame a retailer for not wanting to talk about less jobs, so they failed. The truth is that retail is a cut-throat world where only the strongest survive (although I am constantly amazed by how long the weak manage to hold on), so AI must be pushing at an open door marked Process Optimisation and Data, two areas that retailers need to fix if they are to survive competition and the unreasonable consumer.

Retail is all processes and data now; build the processes that exploit the data quickly, accurately responsively and repetitively, and you have a retail business fit for the future, regardless of short-term considerations about channel format, function, size and mix. In this world of perfect efficiency, it is the humans that provide the difference through service and engagement.

So, we should embrace AI because it, and other technologies, enable everyone to focus on the ‘fewer and better’ retail jobs, first raised by Helen Dickinson at the BRC in March 2016. There will be two types of jobs; those spawned by AI itself; we already know that there is a world shortage of data scientists but I am talking about all the jobs further downstream, working every day in retail. And then there are the other jobs that have been mooted but possibly not well described – in store, helping customers, keeping stores looking bright and tidy, fetching and carrying – all the jobs staff would like to do more of if they weren’t too busy running admin tasks.

Will there be fewer of them in store? Possibly, but will the total number of people employed in retail fall dramatically over the next five years? I don’t think so, and that’s because I cannot see that AI is sitting at the top of retailer’s investment decision pile at the moment. There have been a handful of notable exceptions where AI has been deployed to tackle the big challenges, such as replenishment, but other examples have tended to be tactical and peripheral. And of course, most retailers are not ready for AI because their current processes, data structures and politics are all against it.

That will change of course, but let’s take the time to plan it right. Better first to think about what kind of retail industry we need to serve the new consumer, otherwise AI will simply be used to patch up what is already beyond repair. For instance, AI should be able to help a retailer determine the best location and best forward motion for every product at every moment, but will the retailer adjust its own structure to enable AI to deliver big. And, is the current location planning model based on false assumptions about how customers shop now and in the future?