Black Friday 2025: Digital Dominates, And AI Takes Deeper Root

Black Friday 2025 has come and gone, giving us our first glimpse into holiday period spending, as well as a more concrete look at the shape and scope of the impact that AI is making. 

The numbers, as they are, appear solid rather than spectacular.  Adobe’s tracking placed US digital spend at $11.8 billion (USD) for the day, ahead of its own expectations. Mastercard’s read on overall retail produced a modest year-on-year lift, with the strongest contribution coming from e-commerce. All of this reveals a season where shoppers are participating, but doing so on their own terms, rather than being beholden to some of the behavioural shaping that’s been successful in the past.

A week ago we wrote about that atmosphere of caution, noting that the economic backdrop was uncomfortable enough to encourage planning, on consumers’ parts, without suppressing demand altogether. In that analysis, we also wrote about the new wave of AI-supported shopping features that had only recently begun appearing across major platforms, raising the question of whether any of those tools would matter once the first true test of the holiday season arrived. As it turns out, Black Friday did not deliver a definitive answer, although it did give us a better sense of how these tools are being rolled out and adopted – or, perhaps more accurately, how shoppers are fitting them into the habits they already have.

Across the reporting, a consistent picture emerged of shoppers who approached the weekend with a level of preparation that has become somewhat characteristic of the past two years, where lists were drawn up well in advance and browsing happened early, with many people waiting for reductions to settle before committing. That created the impression of a promotional season that unfolded slowly rather than in a single burst of activity. Gone are the days when stampedes of people would fight each other to get through store shutters on opening in the hopes of grabbing a discounted flat screen TV.  The promotional season has widened enough that Black Friday functions less as an event and more as the finish line in a longer run of November activity.

That rhythm naturally favours digital channels because they provide an environment where shoppers can be more deliberate, which helps when budgets are tight. You can compare five retailers in the time it takes to cross the street to one. You can check stock, colour, price history, and delivery with almost zero effort. You can do it on your sofa, or in a queue, or in the five minutes between meetings, and you can do it all from the convenience of your phone, which continues to dominate how people browse and shop online (According to Adobe). Physical stores have their charms, and they will be back in other parts of the cycle, but this particular weekend rewarded the channel built to support careful decision-making. 

Which brings us to the question of AI and how much it actually contributed. A few outlets were quick to declare that AI “has changed the way we shop,” which is a bold conclusion to reach when what really happened is likely far more grounded. Yes, usage increased, up over 800% according to Adobe. Yes, more shoppers interacted with helpers, prompts, or comparison tools. Yes, retailers saw more traffic moving through AI-enhanced sequences. But the leap from “usage is up” to “behaviour has changed” is enormous, and nothing in the weekend data bridges that gap. 

Where the assistants were placed at the right moment, they pushed conversion in a way retailers felt, though the underlying behaviour stayed largely the same. Essentially, the tools helped people do what they were doing already. They made comparisons easier, but not different. They made discovery slightly easier, but not in a transformative way. If anything, Black Friday showed how well the AI pieces can blend into the background, which is probably the mode in which they’ll offer the most value. For now, as we’re continually told, the agents are coming, just not this year. 

The part of the story that potentially could hold greater relevance for fashion sits away from discovery and closer to the financial realities of the category, because apparel produced strong early numbers (as it usually does during this period) yet the real impact on retailers will only come into view once returns begin to move through the system. 

That side of the cycle doesn’t show up in the first wave of reporting, but it tends to have a very meaningful impact on the picture once it arrives. Multi-size ordering is entrenched behaviour in online apparel. People buy a range knowing they won’t keep everything and then send back anything that doesn’t work. The volume is significant enough that it makes a real dent on the apparent success of the weekend once everything has been processed, and the reality is that it’s one of the structural costs of digital growth.

The sheer scale of those returns is significant, representing billions of dollars of inventory moving back through warehouses. This is the arena where AI might make a more meaningful difference. Returns are baked into online fashion, and not all of them are fixable, because people will always order more than they intend to keep. But the proportion driven by fit and sizing is not trivial. If AI can tighten that up, even modestly, the financial upside for retailers could potentially outweigh anything the new shopping tools have contributed so far.

Physical retail, by contrast, delivered a softer set of results. Some categories held up better than others (apparel dipped only slightly) but overall the decline across regions and segments makes clear that store visits are no longer the headline attraction they once were, and that the event has moved a long way from the door-crashing scenes that used to define it.

A separate development, only tangentially connected to the weekend but relevant in light of how dependent Black Friday has become on digital infrastructure, came from Lush, which reiterated its position on stepping away from major technology platforms such as Amazon, Google and Meta, and encouraged other retailers to consider similar moves, a stance that stands in interesting contrast to a weekend where the tools built by those companies underpinned almost every aspect of discovery, comparison and checkout, although we raise this point not to claim contradiction but as a reminder that the industry still leans heavily on infrastructure controlled by a small group of firms, even as some retailers look for ways to build more independence from them.

Black Friday didn’t bring any big reversals, then. It showed shoppers making careful choices, and it showed digital pulling most of the weight. AI was present but not pivotal, though it clearly made inroads in places where retailers deployed it well. Apparel moved well but will likely edge back again once the returns land. If there is a lesson, it’s probably that the work retailers put into shoring up the digital experience continues to matter more and more, particularly in a year like this one – and almost certainly in a year like the next one that’s just around the corner.