Artificial intelligence is increasingly being used to process big data and make smart decisions. It can be used to analyse online shoppers’ past behaviour, detect fraudulent purchases, and optimise product pricing. AI helps online business owners to better understand their customers. That opening paragraph was written by an AI. I went to copy.ai (which is built on GPT-3, arguably the best commercially available AI engine), asked it to write an introduction for a blog called “Artificial
ial intelligence is changing the way retail is run”, and that’s what it spat out in two minutes. No tricks. Prof Scott Galloway (of “The Four Horsemen” fame) included AI as one of the eight core variables that contribute to trillion dollar companies. While the introduction to this article is cool, much of the real business impact is the stuff which isn’t as in your face, and it’s probably more advanced than you realise. Google and Facebook are largely AI machines. On the ad side, both are pushing to give more control to their AI (and out of your hands). Look at Google’s Smart Shopping, smart bidding, and UAC products which are AI black boxes, money in and (hopefully) results out. Facebook’s moves to limit the number of ads per account is to improve their AI. The world of media buying – both digital and analogue – will be increasingly an AI game and the future of media agencies will be interesting. Copywriting and ad creative are ripe for AI disruption as well. Amazon already uses AI to guide their delivery driver routes, and I know one Aussie retailer doing something similar. What about staff rostering? The order items that are picked in, warehouse layouts, and more are based on tomorrow’s data, not yesterday’s. There’s a lot of tech already but it’ll only get more intelligent. And in a world of AI-driven drones and self-driving cars, how far off are we where a customer can get an order in hours without a single human touching it? It’s interesting in our no-touch Covid times. Product isn’t immune. Trend spotting, ordering quantities, seasonal forecasting – and maybe even design – can all be enhanced or even run by AI. I’ve already seen AI solutions for these. Visual merchandising is ripe for disruption, both online and instore. I’ve experimented, more than once with some AI-driven visual merchandising technology. Ditto for planograms. How about catalogue photos, the bane of fashion retailers? There are already AI solutions for them and I’ve played with them – they are good. You may well have already seen them without knowing. Chat bots on websites are pretty clunky today, but they are getting better – replika.ai is a conversational “caring” bot partly run by GPT-3. Chat bot equivalents of this are in development and they’ll only get better. There are real world equivalents – will shop assistants be a premium offering? Would you rather an ever present helpful virtual assistant or an actual human who you can never find? When you talk to people about the inevitable job disruption of AI, invariably it’s never their job that will be replaced, always someone else’s. This may be an underestimation or misunderstanding of the impact the next decade will have. There’s already early AI-driven UX design tools. Better than a human? Definitely not (at least, not yet). There’s an ethics and disruption discussion there I’m consciously skipping over in this piece. How do we respond as retailers? For a start, AI loves data. The better your data is, and the more you have, the better positioned you are. Tools have been using AI for a few years to communicate to customers on their predicted future behaviour, but they need great data to do that. Of course, online advertisers are data machines, all to feed their AI engines. For smaller retailers, look at what your vendors are doing and how you can embrace that. There’s an opportunity for you to be agile and adapt quickly. There are some amazing startups popping up all over the place, and small companies can pivot fast. Larger retailers, look at investing and changing your processes to take advantage of new technologies. I know one Aussie retailer with inhouse-built AI tools to enhance product discovery. There are people working on predicting return rates. What else can you do? Most importantly, don’t dismiss it. It’s not totally transforming the world right now, but it’ll sneak up and one day you’ll realise you’re left behind. Oh. And watch out for killer robots.