.png)
At a Glance
- +80% revenue per visitor - Colourful Rebel (pending approval)
- +47.7% average order value - Colourful Rebel (pending approval)
- +17.7% engagement rate - Anna + Nina
- +15% revenue per visitor - POM Amsterdam
The clients
Creating and curating eye-catching jewellery, homeware, and clothing, Anna + Nina passionately inspires the world around them through their exquisite and unique collections of carefully crafted and consciously curated products, adding a touch of colour and vibrance for those who dare to dream.
POM Amsterdam is a colour-forward fashion and lifestyle brand built on hand-painted prints, playful design, and carefully selected materials. What sets them apart is the quality and colour of their fabrics.
Colourful Rebel is a vibrant, expressive, and deeply visual fashion brand with a loyal and engaged community. For their customers, fit and styling are everything.
Three different brands, same problem: great static images, but no video.
The challenge
Most fashion and lifestyle brands invest seriously in product photography. The shoot happens, the images go live, and the products look great. But on the PDP - where the buying decision actually happens - it's all static. No video. No movement. No way to really get a feel for the product.
An image shows what a product looks like, but it can't show movement, texture, or how a product actually fits, looks and feels in real life. For fashion and lifestyle brands, where customers are buying confidence as much as a product, that's a problem. And it shows up in the numbers - lower conversion rates and higher returns.
The data is clear: Video outperforms static on product pages, paid social, and email. But for most brands, producing video for every product has felt out of reach. Too expensive, too slow, too hard to scale.
Our approach
At GoodKarma, we have spent the past year building an AI creative production capability - not as an experiment, but as a service with a clear methodology and consistent results across clients.
We start every project by figuring out what the brand actually needs. Where are customers dropping off? What's missing from the product page that would give them confidence to buy? Is it movement, fit, texture, detail? Once we understand the real gap, we look at what assets the brand has to work with, what rights they hold (including model consent and creation rights) and write a brief from there. The tech serves the brief. Not the other way around.
We also take the rights position seriously - both legally and ethically. Before any AI production begins, we require that the appropriate rights are in place: shoot contracts, model consent for AI use of likeness, and photographer creation rights. The photographers, models, and creative teams whose work is what we're building on, and we respect that. We won't start a project until those permissions are confirmed.
From the audit, it becomes clear what kind of content we’ll be generating. We currently work across four types:
- Product animation takes existing product photography - typically studio or silo shots - and gives it movement. Texture, detail, sparkle, and depth become visible at scroll speed in a way static images rarely achieve.
- AI model generation and animation builds a brand-aligned model from scratch using AI when no existing model assets are available. The model is briefed in detail (look, body type, styling, poses, and overall brand fit), all decided together with the client until it feels like a natural part of the brand.
- Campaign video adapts existing brand or campaign imagery into video-ready creatives for paid social, email, or organic channels. Performance-first, fast to produce, no new photography required. Rights review are needed here, too.
Three Brands. Three Briefs. Outstanding Results.
Anna + Nina - Product Animation for PDP
Bringing high-value jewelry to life from existing product photography
What we found
Anna + Nina's Vintage Collection and 14k gold jewelry were generating strong, consistent traffic on PDPs - but the assets were limited and it was hard to convey the detail and quality of the product through static images. When we looked at the data and how people were actually moving through the site, it was pretty clear what was happening: customers were landing on high-value product pages and leaving without engaging. Not because they weren't interested, but because the photos weren't giving them enough to buy with confidence.
The craftsmanship, the sparkle of a stone, the weight and texture of a piece - that's what makes one piece different from another, and justifies the price point. And none of it comes through in a photo. The confidence gap between image and product was costing conversions.
What we did
We introduced a short, autoplaying AI-generated product video as the second image on a selection of high-value PDPs - Vintage Collection and 14k gold - keeping it out of the PLP view while ensuring it was the first thing a visitor saw on the product page itself. Static images stayed in the gallery, but he video was the entry point.
On Rights: Product-only animation using brand-owned photography. No model involved, no shoot contract or third-party consents required.
The results
- +17.7% product engagement (74.3% vs 63.1%)
- +8.54% add-to-cart rate
- +15.3% checkout starts
The video did what we hoped: closed the confidence gap. More users engaged with the products, more added to cart, and more made it to checkout.
What Anna + Nina can now do
AI video is now part of the standard launch process for every new jewelery higher-value piece. We've set up a recurring production cadence with a clear priority framework - becoming the new standard for how products go live.
POM Amsterdam - AI Model Generation and Animation for PDP
Building a brand-aligned model from scratch when no assets existed
What we found
POM Amsterdam needed model video on their product pages. Research consistently shows that video improves product understanding in fashion - especially when it comes to movement, fit, and texture. Those are the things that matter most when you're buying clothes online, and they're exactly what static images struggle to communicate.
The problem was they didn't have any model video for their current collections. Shooting it traditionally would mean casting, styling, a full production day, and then all the editing afterwards - weeks of lead time and significant budget for something they hadn't tested on their PDPs yet. Before committing to all that, we wanted to prove the concept first.
What we did
We generated a brand-aligned model from scratch using AI. This is not a technical process that runs on autopilot - it is a creative brief. Representation, styling, lighting, aesthetic alignment with the brand's existing photography and brand guidelines - all specified upfront, in close collaboration with the POM Amsterdam team, before a single frame was produced. These decisions reflect brand values, and we treat them as such.
From there, we generated multiple models with different looks, movements, and behaviours - and shared them with the brand. It's essentially a digital casting process. We refined and iterated together with the POM Amsterdam team until we landed on the right model - one that felt like a natural extension of the brand’s visual identity.
The selected model was then animated into 5–10 second autoplay videos and deployed as the second visual on top-traffic PDPs. No real person's likeness. No agency. No traditional shoot.
On rights: The model is entirely AI-generated - no real person involved. A creative brief and explicit client sign-off on representation were required before production began.
The results
- +7.44% conversion rate
- +15% revenue per visitor (€8.24 vs €7.17)
- +25% average products per visitor
The numbers speak for themselves. Visitors who engaged with the video converted at a higher rate, added more products to their carts and spent significantly more per visit. The video gave people the confidence to buy, and to buy more.
Based on these results, we rolled it out across all new season products and are now testing placement, format, and load speed to push it further.
What POM Amsterdam can now do
New product launches can go live with model video from day one, without scheduling a shoot. This isn't a replacement for professional model photography or video. But a great alternative for when a shoot isn't an option - whether that's budget or timing.
Colourful Rebel - Existing Model Animation for PDP
Turning an underused shoot archive into an active video library
What we found
Colourful Rebel had great model photography from previous shoots, but no video. The images were strong, the styling was on point, and the creative investment had already been made - but there was untapped potential to make them work harder. We wanted to see if we could give those photos movement and get them onto PDPs as video, without a single day of new production.
What we did
We identified key product categories and took the existing model photography for those products to animate it into short looping videos. The goal was to bring movement, depth, and energy to imagery that was already on-brand - without any new model, new shoot, or new styling. Just the existing photos in a completely new format.
From there, we tested different animation styles and movement behaviours to find what felt natural for each shot - helping bring the product details, fit, and styling forward in a way static images never could. The final videos went straight onto PDPs as the second visual.
On Rights: Three things had to be confirmed before work began: that the original shoot contract permitted derivative or AI-generated works, that the model had consented to AI use of their likeness, and that the photographer's creation rights allowed it.
The results
- +26.2% conversion rate (5.15% vs 4.08%)
- +80% revenue per visitor (€9.38 vs €5.21)
- +47.7% average order value (€173.63 vs €117.52)
- +78.5% total revenue generated during the test period
Visitors who saw the animated model content converted more, spent more per order, and generated significantly more revenue per visit. The video gave people a better sense of fit and styling, and that confidence showed up across every metric.
What Colourful Rebel can now do
Every model photo in their library is now a potential video. New shoots can be animated as part of the standard post-production process, so every shoot delivers more than just stills.
Which Use Case Is Yours
Not sure which approach is right for your brand? These three questions will help you figure it out.
What assets do you have?
- Product photography only → product animation.
- Existing approved model photography → model animation, rights review first.
- No model assets at all or no animation rights → AI model generation.
What rights do you hold?
- Brand-owned product photography → straightforward, all options open.
- Model photography → shoot contract and model consent for AI use must both be confirmed before any work begins.
- Commissioned creative for paid social → confirm usage rights cover paid placements specifically.
What channel are you feeding?
- PDP or PLP → all three use cases apply.
- Paid social or CRM → the same assets can be adapted for campaign placements without additional production, once on-site rights and usage scope are confirmed.
Key Takeaways
Your existing creative can work harder than it currently is.
Every product image and approved model photo your brand owns is a potential video asset - not only on your PDPs, but also in paid social and CRM. And when the right assets don't exist yet, we can create them. The shift is straightforward: from static to video, across every touchpoint that matters.
The content gap is no longer a budget problem, it's a process problem.
The brands winning on video aren't outspending everyone else - they're outproducing them: more formats, more channels, more touchpoints, from the same underlying creative investment. We think this is where the next real performance gains in fashion e-commerce are coming from, and the brands that build it into their process now will be hard to catch.
Rights are not the small print, they are the brief.Model consent, creation rights, and usage scope need to be confirmed at the start of every project, before a single asset is produced. It protects the brand, respects the people behind the original work, and means nothing ever has to be pulled.
Closing
Three brands came to us with the same gap: great photography, no video, and no realistic way to produce it at scale. Instead of working around that limitation, they decided to test a different approach, and the results speak for themselves.
This is what happens when you combine AI creative production with real brand understanding, a clear methodology, and respect for the original creative work. We didn't replace what these brands had built, we made it work harder: more formats, more touchpoints and more revenue per visitor - from the same creative investment.
That's the GoodKarma approach: brand and performance, together.
If you want to know what AI Video Generation could do for your brand, get in touch.
Share on:
.png)
