Published 2026-05-15 · Part 1 of 6 — ECD Munich post-event communication series
The Inbox Every CMO Opened This Quarter
On May 12th 2026, 600+ e-commerce leaders gathered at ECD Munich. Same week, McKinsey published State of Fashion 2026. Same quarter, every fashion CMO received the same vendor pitch: "Cut your content costs 10×. Replace your photoshoots. Scale your catalogue instantly."
The pressure to say yes is real. Your CFO wants the savings. Your marketplace team needs 4× the SKUs you produced two years ago. Three competitors already announced their AI partnership.
But the decision most brands are about to make will widen the gap they're trying to close.
Kamil Czaja opened his ECD masterclass with a single line: "Demand is down. AI promises are everywhere. The truth behind the desk." What followed was a 45-minute dissection of the collision between market reality, customer behavior, and the AI vendor pitch that landed in your inbox last Tuesday.
This is post one of six This is the first of six posts on what 120 fashion brands and 600 e-commerce leaders learned in Munich in Munich. If you're evaluating an AI content vendor this quarter, read this first.
The Market Is Pressing You Harder Than Ever
Fashion demand is the lowest it has been in a decade. McKinsey State of Fashion 2025 projects +1% EU fashion market growth for 2025. Pre-pandemic, that number hovered near 5%.
Return rates climbed +3.0pp across fashion between 2022 and 2024, according to Coresight Research 2025. Customer dwell time on product detail pages dropped 22% versus 2022, per Contentsquare Digital Experience Benchmark 2025.
Kamil's line from the stage: "You can't grow demand. You can win bigger share of every visit."
That shift in framing changes everything. When the market grows 5% per year, you can afford inefficiency. When the market grows 1%, every abandoned cart is revenue you needed. Every return is margin you lost. Every customer who bounces from your PDP after 8 seconds just handed their wallet to a competitor.
The only lever you control is conversion. The only surface where conversion happens is the product detail page. The only input the customer evaluates in those 8 seconds is your content.
The Customer Changed Faster Than Your Catalogue
Gen Z and Alpha will drive 40% of fashion spend by 2030. Per capita spend in this segment grows +7% annually, according to McKinsey 2025. 41% of Gen Z already use AI chat tools weekly to shop fashion, per BCG 2025.
Their journey starts in ChatGPT. They ask for "linen blazer under €200, structured fit, neutral palette." The AI returns eight options. Three of those are your competitors. One is you.
They click through to your PDP. Now the decision happens. The decision happens on your page, not in the chat.
59\.9% of online shoppers question content authenticity, according to Accenture Life Trends 2025. 71% are suspicious of images that look AI-generated, per Salsify 2025. 78% don't consider AI imagery authentic, according to Getty VisualGPS 2025.
Here's the paradox Kamil landed in Munich: "The same customer who uses AI to shop refuses to be sold to by AI."
They want AI to help them find the product. They want real photography to prove the product is worth buying. If your PDP shows a model with six fingers or a hem that defies physics, they close the tab. If your packshot looks like it was rendered in Blender, they assume the fabric quality is fake too.
44% of shoppers deliberately avoid brands they suspect of using AI-generated imagery, according to Salsify 2025. Skepticism peaks among 25-40 year olds—the segment with the highest lifetime value.
Anita Balchandani at McKinsey said it plainly in State of Fashion 2026: "If content was king, it's now going be an emperor." Translation: content budget share will grow. Content quality will become the primary differentiator. Brands that treat content as a cost center will lose to brands that treat it as a conversion engine.
Why Your Board Wants to Say Yes to Cheap AI
Four forces make the cheap-AI pitch Four forces make the cheap-AI pitch hard for a CFO to wave away:
1. Cost optimization. Production costs climbed 12-18% since 2021. Demand grew 1%. Board pressure to cut OpEx is intensifying every quarter. A vendor walks in offering to cut content production cost by 90%. That's a €2M annual saving for a mid-sized brand. The CFO hears "fund two other initiatives."
2. Speed expectation. Drop calendars accelerated. Zara went from 5 weeks to 2 weeks. Shein launches 6,000 SKUs daily. Your marketplace team is being asked to triple catalogue depth in six months. Traditional photoshoots can't scale at that pace. AI can generate 500 images overnight.
3. Scale ambition. Marketplace growth requires 4× the SKUs you produced two years ago. Zalando is pulling out of wholesale and shifting to marketplace models. Hundreds of brands just lost the content support they used to get free. Your only option to stay visible on that platform is to produce content yourself. AI is the only tool that can produce at that volume.
4. FOMO. Three competitors announced AI partnerships last quarter. Your VP Digital forwarded you the press release. Doing nothing feels like falling behind. The board asks: "Why aren't we doing this?"
The pressure is real. The forces are rational. The cheap-AI promise is designed to exploit exactly this moment.
The AI Promise Is Real. But Partial.
The vendors are not lying. AI tools do cut nominal production cost by approximately 90% on specific use cases. Speed improvements are measurable. Some workflows are 100% deployable today.
But three things the pitch deck doesn't show you:
Hallucination variance. Internal testing at GoPackshot across 30 days on the same prompt set showed hallucination rates ranging from 14% to 70% depending on the day, the model version, and the product category. Structured garments (blazers, coats) hallucinate less. Draped garments (dresses, skirts) hallucinate more. You don't know which 30% of your catalogue will ship with six-fingered models until after you've replaced your workflow.
Offshore retouching with no brand context. Most AI vendors route post-generation QA to offshore retouching teams. Those teams have never seen your brand book. They don't know your customer. They don't know that your €400 coat is purchased by a 38-year-old banker in Copenhagen who will return it if the lapel looks wrong. They fix the six fingers. They don't fix the trust gap.
The trust gap widens, not narrows. 71% of shoppers are already suspicious. If you deploy AI-generated imagery across your PDP grid without a quality floor, you're handing them proof that their suspicion was justified. First return costs you margin. Second return costs you the customer.
Daniel Kahneman won the Nobel Prize in 2002 for work published in Thinking, Fast and Slow. His framework explains why this matters:
System 1 is fast, unconscious, emotional. It drives 95% of decisions. Photos work here. PDP loads. Customer asks: "Do I like it? Is it real? Is the price fair?" If yes, they buy. This is where the sale happens.
System 2 is slow, logical, effortful. It drives 5% of decisions. Reason wakes up when the delivery box opens. Customer compares expectation versus reality. If there's a gap, they notice. Refund ticket gets filed. Trust in the brand is lost. This is where the return happens.
The photo proves the price. 87% of shoppers pay more for a brand they trust, according to Salsify 2025. 57% would pay a premium if products match the images, per Nfinite (N=1,000+ US consumers).
Cheap AI optimizes for System 1 speed. It breaks System 2 trust. You get the sale. You lose the margin. You lose the customer.
What Adrian Heard on the Floor in Munich
Kamil's keynote framed the tension. The hallway conversations revealed how brands are responding.
One major fashion retailer (name withheld) shared an internal metric: average time from sample arrival to product online is 65 days. 70% of products never make it online at all. Missing content stops launches. A single absent packshot can block the entire product from going live.
The CEO of a major European marketplace told a group at the booth: "We have AI that does everything itself." No human review. No fashion experts in the loop. Full automation from upload to publish.
Zalando's wholesale pullback came up in four separate conversations. Hundreds of brands that relied on Zalando to produce their content for marketplace listings just lost that support. Their options: build in-house production, outsource to a studio, or deploy cheap AI. Most are choosing option three because options one and two take six months to spin up.
The strongest reaction at the GoPackshot booth came from a framework Kamil previewed: "Send us your packshot and model shot. We'll generate context images." Girl in your shirt on a bike. At the beach. In motion. Real product. Real base photography. AI-extended context. That use case landed because it keeps the trust anchor (real product photo) and uses AI where hallucination risk is lowest (background, scene, secondary elements).
What This Means for You
If you're evaluating an AI content vendor this quarter, you're not wrong to explore it. The pressure is real. The promises are partially true. The cost savings are material.
But if you deploy cheap AI across your entire PDP grid without understanding where hallucination breaks trust, you will:
1. Increase return rates (because AI-generated images set expectations the physical product cannot meet)
2. Decrease repeat purchase rates (because customers who return twice don't come back)
3. Lower conversion rates over time (because 71% of shoppers are already suspicious, and you just proved them right)
The decision you make this quarter will either close the trust gap or widen it.
This is post one of six This is post one of six on the ECD Munich insights and the framework 120 fashion brands are using to evaluate AI without breaking trust and the framework 120 fashion brands are using to evaluate AI without breaking trust.
Next week: The Trust Gap Deep Dive: Why 44% of Shoppers Avoid Brands Using AI, and the 15 Use Cases Where AI Actually Works.
Subscribe to the series at gopackshot.com/blog or reach out directly if you're evaluating a vendor and want the full framework before you sign: contact@gopackshot.com.
