Shopping, artificial intelligence sees you. And even buys your clothes

Here, too, the answer may come from AI. With the emergence of increasingly sophisticated Virtual Try-On apps. We upload a photo of ourselves and immediately see, on screen, what that outfit looks like on us. Google integrated the function on AI Mode this year. An alternative is the app Doppl, by Google Labs, which also generates animations of our body in that garment. Both services are only available to American users, for now. These apps may recall the concept behind augmented reality systems that already allow us to see what a large object (a carpet, a vase...) looks like, virtually, in our room. They are also on Amazon in Italy. With clothes and human bodies, however, it is more complex, hence the need to make use of artificial intelligence. And that's just the beginning: generative AI now breaks the boundaries of fantasy. AI Mode also makes it possible to describe a dress in words, to see it generated in pictures and then have links to sites where to buy a similar one.

But if the goal is an easy and straightforward shopping experience, why not automate it as much as possible? This is the principle behind AI shopping agents. Many AI-OpenAI players, Google, Perplexity, Microsoft - and the American retail giant Walmart - are working on browser agents that search for products, compare them and fill e-commerce shopping carts instead of the user. All they leave us with is the final click to purchase. At the moment it is too dangerous to automate even the payment: the AI could 'hallucinate' and stuff nonsense into the cart. It has already happened in some field trials for these agents. Good then that there is still a final human verification.

Moreover, this is not the only obstacle to the take-off of agents: at the moment they are slow and consume a lot of computing resources in order to function (hence they are wasteful for suppliers and, cascading, users).

More useful perhaps is to look at the pragmatic ways in which online marketers are already now embracing generative AI, primarily with chatbots for customer service and conversational search. The most innovative ones will integrate more sophisticated AI functions that are clearly useful for the end user. For instance, Google already offers an API (programming interface) to integrate Virtual Try On into its apps or e-commerce sites.