Three stories this week show the same shift playing out at different scales: AI is no longer a cost experiment — it’s showing up in profit forecasts, search traffic, and the runway itself.
1. Zalando’s AI Investment Pays Off in Real Numbers
On March 12, Zalando reported its 2025 full-year results and made a clear case that AI investment translates to financial results.
Active customers grew from 51.8 million to 62 million.
Gross merchandise volume hit €17.56 billion — up 14.7% year on year.
The company projects adjusted EBIT of €660–740 million for 2026,
compared to €591 million in 2025.
Shares jumped 7% on the day.
What drove it? The company pointed directly to AI.
Using AI-generated product images,
Zalando published 70% more advertising content without a proportional increase in creative headcount.
The company also launched an AI virtual try-on feature —
the kind that helps shoppers select the right size before buying —
which reduced size-related returns,
one of the biggest cost levers in online fashion retail.
The story matters because it is not a pilot program or a press release about future potential.
It is a quarterly earnings report with measurable numbers attached.
The takeaway: AI image generation that saves cost is table stakes.
The companies pulling ahead are using it to produce more content at the same cost —
higher volume, faster cycle, same team.
2. Which Brands Are Winning in AI Search — and Why
A March 3 BoF Insights study (conducted with Quilt.AI) looked at how ChatGPT,
Gemini, and Claude recommend fashion brands to shoppers —
and found that the gap between winners and everyone else is already wide.
The numbers behind the stakes: referral traffic from AI platforms to US retailers grew 760% year on year in November 2025, according to Adobe data.
Zara gets 15% of its referral traffic from AI; H&M gets 8%.
Shoppers arriving via AI are also more likely to complete a purchase than those arriving from other channels.
Among high street brands, Uniqlo, Zara,
and H&M receive the most frequent AI mentions.
Skims performs well as a category specialist,
showing that you do not have to be a giant to win here.
The research identifies two paths to AI discoverability: broad usefulness (generalists like Zara and Uniqlo, which AI recommends as safe, reliable choices) and sharp category definition (specialists like Skims, which AI recommends when someone asks a specific style question).
Both paths require the same tactical work: updating product descriptions with specific lifestyle contexts, improving imagery, and consistently repeating brand values across all digital platforms so AI systems have enough signal to retrieve and recommend your brand.
The takeaway: If your product listings use generic descriptions (“blue denim jacket, size M”), AI has nothing to work with.
The brands winning AI search are the ones that write for real use cases —
“weekend camping look,” “minimalist office layering” — not SKUs.
3. Fashion AI Gets Its Own Expo — and a Debate to Match
The first dedicated Fashion AI Expo ran during Paris Fashion Week on March 6.
The event brought together AI tool developers, fashion brands, buyers,
and investors in one place — an acknowledgment that fashion’s AI conversation has grown big enough to need its own stage.
Among the companies showcased,
Heuritech stood out for the scale of its operation: the trend analytics firm processes 3 million social images per day to track what is actually appearing on social platforms versus what designers are forecasting.
The AI-generated fashion market was valued at $2.14 billion in 2024,
and the expo positioned itself as a recurring meeting point as that market grows.
The reaction was not unanimous.
Investors welcomed the concentrated deal flow.
Creative directors pushed back,
with some arguing that automated trend signals risk making collections look more alike — that if every brand runs the same AI trend data, the results converge.
It is a real tension, and it is worth tracking as these tools mature.
Paris Fashion Week itself generated $629 million in measured media impact that week.
The expo’s co-location was deliberate —
fashion’s creative center and its AI tools in the same room, at the same time.
The takeaway: The expo debate — efficiency versus aesthetic differentiation —
is the same decision every brand team faces when adopting AI.
The brands that use AI to do more original work will diverge from those that use it to copy what is already trending.
What This Means for Your Brand
Three things to act on this week:
-
Update your product descriptions for AI search. Zalando and Zara invest in rich, context-specific product content.
Generic descriptions do not surface in AI recommendations.
Start with your top 20 SKUs and rewrite them with lifestyle context: who wears it,
when, and with what. -
Track AI referral traffic in your analytics. If you are not already segmenting AI referral traffic separately in your analytics, set it up.
Adobe data shows 760% year-on-year growth in the US —
your market may not be there yet, but the trend is clear and worth measuring early. -
Think about AI image volume,
not just quality. Zalando’s 70% content increase was not about making one perfect image.
It was about producing enough variations to run more tests, serve more placements,
and iterate faster.
AI image tools — including lookbook generators —
are most valuable when you use the volume advantage they offer.
Sources: Zalando Eyes Higher 2026 Profit as AI Drives Productivity —
RTE Business, Which High Street Fashion Brands Are Winning the AI Discovery Battle?
— Business of Fashion, Fashion AI Expo Reveals March 6, 2026 Program — Glass Almanac