Where have we got to with AI?

There is little doubt that you’ve heard the term ‘AI’ 25 times already today. It gets bandied around so much it can sometimes feel like it is losing its meaning, if it ever even had a concrete sense of one in the first place. So what does it mean, in an ecommerce context, and where are we with it?

At its most basic level, AI means ‘artificial intelligence’, which specifically refers to a kind of technology that is advanced enough to perform various tasks autonomously. This is different from, for example, a traditional set of traffic lights which appear to operate on their own, as they are actually just set on a timer and will continue to change in the middle of the night when there is no traffic. AI systems use data to inform decisions; so AI traffic lights might change their pattern through assessing criteria such as the load of cars coming in each direction, volume and type of pedestrian wanting to cross or emergency service vehicles trying to get through.

Transposing that concept over to business sectors such as retail, the promise in terms of productivity and efficiency is clear to see. For ecommerce sites, where the typical conversion rate hovers around the 3% mark, we can start to imagine AI systems combining disparate datasets together to create experiences for site visitors that are highly relevant and engaging on an individual level. Instead of that 3% conversion rate, could we dare to dream of it increasing to 10%, or 20%? The fact is, even a fairly modest increase to 5-6% would be transformational for the ecommerce industry.

So, are we there yet?

In a word, no. Some of the biggest changes we’ve seen in technology over the past 15 years or so have made a difference and pushed growth levels up for ecommerce, particularly the introduction of tablets followed by smartphones, as they created new contexts in which people could engage with retailers’ product ranges. It doesn’t always work; concepts like the metaverse and voice devices have been introduced without even a ripple of impact.

AI though feels like it has potential that far outweighs what these other technologies promised. This is because it is not (or, will not be) a separate thing operating by itself, but a generalist concept that seeps into every other existing technology; the proverbial rising tide that floats all ships (though some ships clearly remain bigger than others…). So when is the great wave going to arrive?

There is good news and bad news on that front. Let’s start with the bad. Every tech-provider is loudly proclaiming themselves to be ‘AI’ now, which suggests that the platforms retailers use widely already offer this functionality; if this is reliable information, the evidence should be in full flow. This, in turn, should be creating a productivity boon where processes get faster and slicker. Check the macroeconomic news though; I’m not sure we are seeing real evidence of this quite yet.

This may be a reflection of the hype maybe getting a bit ahead of the reality, but there is good news on this front too. In 2024, whenever I asked groups of retailers what they were doing with AI, there tended to be a lot of head-scratching and a general reluctance to claim that they were getting anywhere. Things seemed to be moving so fast that retailers couldn’t quite keep up with it.

Fast-forward to 2025 though and I note a sea-change in response. Now whenever I ask retailers that same question people seem a lot more confident in saying they are engaging with it – large companies and small alike. In some cases it is just using generative solutions such as Chat-GPT, but there is some real experimenting going on too. For example: consensus seems to be that model / product-generation – which could speed up photography sessions massively and reduce costs – is not quite there yet, but it isn’t that far off either.

Are we really on the cusp of serious automation and the provision of genuine one-to-one personalisation at scale? Perhaps we aren’t really in the midst of the great AI revolution just yet, but the progress and switch in attitudes is palpable. The speed at which things are developing means 2026 might be a year of real progress toward it…could that productivity boon start to peek out from behind its cloud next year?