And they should be. There are big promises, big opportunities, and an ever-growing list of things AI can do for e-com brands.
But that's also part of the problem. There are almost too many things you can do with AI…
Some of the use cases are super flashy and making headlines: AI models, Digital Twins, campaign image to video, and entirely generated campaigns. Naturally, those are often the places brands look first.
But after talking to a number of brands about implementing Gen AI workflows, we've noticed something interesting.
The most exciting AI use case isn't always the best place to start.
In fact, some of the most practical opportunities we've seen are much less obvious.
Relighting existing photography instead of reshooting it. Turning on-model photography into packshots. Creating better B2B and wholesale imagery. Adding motion to existing product photography.
They might not generate quite as many headlines, but they can solve very real studio problems, and in some cases, they're much easier to introduce into an existing workflow.
So again, we're saying it: Where does AI actually make sense?
That's what our AI use case matrix is designed to help answer.
Because building an AI business case isn't just about proving ROI. You also need to understand what introducing a particular workflow means for your organization.
Does legal need to be involved? Do you have the necessary model rights? Does the studio need to change how it shoots? Who art directs the output? Do existing roles need to change? And how much creative oversight will the workflow require?
Suddenly, "easy" and "difficult" become a little harder to define.
That's why our framework looks at AI use cases through four practical lenses: legal considerations, workflow impact, creative complexity, and business value.
From there, we can start thinking about AI adoption in three stages: Crawl. Walk. Run.
We'll admit, this is a simplified version of the matrix. But we think it's a useful starting point for brands — not only for evaluating these use cases, but for thinking about any AI opportunity through the same four lenses: legal, workflow, creative, and impact.
The AI Use Case Matrix
| AI Use Case | Legal | Workflow | Creative | Impact |
|---|---|---|---|---|
| Non-Public / B2B Content | Easy | Easy | Easy | Strategy |
| Relighting | Easy | Easy | Easy | Faster Scalability |
| Image to 360° Spin | Easy | Easy | Medium | Better Conversion |
| Model to Packshot | Easy | Medium | Medium | Faster Scalability |
| Image to Video | Medium | Easy | Medium | Better Conversion |
| Model to AI Background | Easy | Medium | Medium | Faster Scalability |
| Model Swap | Medium | Medium | Medium | Lower Costs |
| Styled Looks | Medium | Hard | Medium | Better Conversion |
| Packshot to AI Model | Hard | Hard | Hard | Lower Costs |
| AI Campaign | Hard | Hard | Hard | Lower Costs |
Crawl: The Opportunities in Plain Sight
The Crawl stage is the low-hanging fruit, aka the use cases that are relatively easy from a legal, workflow, and creative perspective, but can still mean a lot of time and cost savings for you and your team.
These aren't always the use cases brands think about first. They're the less obvious ones, but that doesn't mean that they can't make a big impact.
Take relighting. A studio can continue shooting as usual, then use AI to experiment with entirely different lighting treatments afterward. Or maybe you already have hundreds of existing assets and want to explore a new brand aesthetic without going back and reshooting everything. AI gives you the flexibility to test that new look, see how it performs, and scale it if it works. Or image-to-360° spin, where existing product imagery can become the starting point for motion without introducing a completely separate production workflow.
And Crawl isn't limited to the examples in this matrix. There are plenty of low-hanging AI use cases we haven't added to our matrix yet. Look at the projects coming up, the gaps in your workflow, or the things you wish you could do more easily. The next great Crawl use case might already be sitting somewhere in your own work.
Walk: When AI Starts Changing the Workflow
Walk is where things start to get a little more complicated.
The use cases in this stage — Model to Packshot, Image to Video, Model to AI Background, and Model Swap — can unlock scalability, conversion, or lower costs. But unlike Crawl, there's usually something more to work through first.
And that "something" isn't the same for every use case.
Model to Packshot is a good example. Instead of shooting both on-model and separate flat or ghost mannequin imagery, brands can use their on-model photography to generate those additional assets. But that doesn't mean one hero image suddenly gives AI everything it needs. You may still need multiple angles or detail shots showing things like hems, cuffs, or parts of the garment that aren't visible. The workflow becomes more efficient, but the studio also has to start thinking differently about what it captures in the first place.
Image to Video brings up a different consideration. The workflow itself can actually be pretty straightforward: you already have the still images, and those become the starting point for video. But translating a still image into motion introduces a lot of new creative decisions. How should the model move? What happens in the scene? How should the camera behave? And how do you translate your brand's existing visual style into the prompts and direction behind the generation?
That's where some creative oversight is still needed. The AI can generate the motion, but someone needs to make sure the result still looks and feels like your brand.
Then there's Model Swap. Technically, this is one of the easier workflows. A brand could continue shooting with a real fit model and then use AI to swap that person for another model. But suddenly the biggest question isn't technical at all. It's about your company's stance on using AI this way, depending on what you swap the model out for — is it a Digital Twin? An AI Model? Do you have the legal backing to make your decision?
That's really what Walk starts to reveal.
An AI workflow can be technically easy and still be complicated to implement.
You might need different source imagery. You might need to talk to legal or your modeling agency. You might need new creative decisions or an internal conversation about where your brand draws the line.
The opportunity is there. Walk is about making sure the rest of the organization is ready to walk with it.
Run: When AI Changes How Content Gets Made
Then we get to Run.
This is where we find some of the use cases that probably come to mind when you think about generative AI in fashion: Styled Looks, Packshot to AI Model, and AI Campaigns.
And yes, these are the ones that tend to get people excited. But they also ask a lot more of the organization behind them.
Take Styled Looks. AI can make it possible to create five, ten, or potentially many more styling variations from the same product, something that would be incredibly difficult to scale on set. But you're also moving from physical styling to digital styling. Someone needs to build the looks, monitor the generations, make adjustments, and decide what's actually good enough to move forward with — or work. Or, you can work with a partner like Pixelz, where much of that generation, review, and refinement is managed as part of the workflow. The skills might already exist within your team, but how they're being used starts to change.
The same applies to Packshot to AI/Synthetic Model. Suddenly, casting can become digital. Pose guides need to be considered. Styling moves from the physical set into a generated workflow. And someone still needs to make sure every image looks and feels like your brand. Many of the creative skills are familiar, the day-to-day process of applying them is what's different.
And then there are AI Campaigns, which bring many of these considerations together. There is a lot more art direction involved, and knowing how to work with the generation itself becomes a skill. When do you regenerate? What should you tweak? When is an image good enough to stop prompting and move into post-production? Again, is there any legal involved with a digital twin? The judgment can start to feel like a skill in itself.
But Run doesn't mean don't start here. In fact, a smaller AI campaign could be a perfectly reasonable way for some brands to test the waters.
That's the point of the matrix.
Run isn't necessarily the destination, and Crawl isn't necessarily the starting point.
It's about understanding what each use case will ask of your organization and deciding where the opportunity is worth the change.
And these are only the use cases we're seeing across the industry right now. There are probably a lot more waiting to be tested, explored, or maybe not even thought of yet.
The matrix can evolve too. Legal, workflow, creative complexity, and business impact are the four lenses we've used here, but your organization might have other questions that matter just as much. Maybe it's data and privacy, technical integration, resourcing, brand risk, or something completely specific to the way your team works.
So when a new use case comes up, bring it back to the matrix. Add a column if you need to. Change the questions. The important part is understanding what the opportunity could give you and what it will ask of your organization to get there.
If you're curious about AI projects or have a use case that doesn't quite fit into this matrix, drop us a line. We're always interested in exploring what else is possible with AI, especially when it starts with a real challenge. Tell us what you're trying to solve, and let's see what we can do.