Six things I just learned about Artificial Intelligence
At our Retail Connections event last night, I realised about 10 minutes in that after six months of monitoring the media and the Internet to get educated about Artificial Intelligence, not only was I none the wiser, I was actually a lot worse off.
Hearing from Nick Lansley from Lansley Consulting, Anusha Couttigane from Kantar Retail, Paul Wilkinson from Tesco Labs, and Katharina Wittgens from Innovationbubble, made me realise with some horror that both the media, and the Internet have gone down the wrong track.
A poll of the retailers in the room revealed that 50% are intending to use AI in their businesses in the next 12 months; 10% already are; and 40% are not yet sure what they should be doing.
- The mainstream media has leapfrogged the present and just wants to talk about a dystopian future where everyone in retail has lost their job. They inevitably focus on the store, which is actually the one place in retail that will probably keep most of the jobs. It is head office where the jobs will go, but no-one yet knows what new jobs will get created once we start to manage data and the decisions it drives, in new ways.
- AI on the Internet is just the tech industry having a conversation with itself. Retailers are either not being included, or are choosing not to take part in the debate. It takes only a few minutes in a room with retailers to realise that, while AI is here to stay, its growth and its contribution are going to take time, and all the hype from the vendors is just going to slow things down.
- Tesco has been automating replenishment for more than 15 years and no mention is ever made of AI. That’s not to say that there is not room for AI to bring further benefits, but vendors that kid themselves that retailers will jump on board just because AI is ‘new,’ are probably mistaken. Retailers will apply the same rigour to their AI researches as they would to any technology.
- The community that loves an acronym is the one that holds back new technology adoption. I get what IoT is, why it’s important, and why we do need to start connecting devices. But analysts, pundits, consultants and academics are just making it harder for retailers to plot their own journey. And so it is with AI to some extent; I get that we need more data scientists, but really, they are not generally the best people to be promoting it, because they get mesmerised by their own cleverness too early in the conversation.
- There are plenty of smart technologies already in place that do not use AI. Store fridges that are monitored continuously to detect fluctuations in electrical current, can be fixed before they actually go wrong, so that sales are not interrupted. It’s practical, it’s affordable and the business case is clear, but it’s not AI.
- AI will not save retailers that have too many broken, missing or disintegrated processes. A bad retailer is a bad retailer and technology will simply automate the mess. In the days when everyone was making money, selling through only one or two channels and dominating a subservient customer, none of this mattered. Today, how many retailers are running modern technology in every area of their business, but wondering why they are not making enough margin or getting growth? Optimising the tech is a part of the answer, but so is processes. And AI can’t fix bad processes.