The problem
- Untapped catalogs. Merchants hold product images but lack time and resources to make them premium and on-brand.
- Expensive shoots. Professional photography is costly and slow to repeat every season.
- Fragmented tools. Existing AI platforms force constant switching and manual re-uploads.
The approach
- Catalog-to-store. Generate premium images and videos from existing products, sync results back without leaving the platform.
- Auto-Config engine. Keeps every output and styling on-brand and on-standard at scale.
- Embedded app. Proven imagery capabilities delivered natively inside Shopify.
Bringing Photogenix to Shopify
The project began as an effort to extend Photogenix beyond Streamoid's CXO suite. Looking at where the capability could create the most value, we identified Shopify as the strongest platform, and a clear gap: independent merchants who need enhanced, on-brand imagery for their storefronts but don't have the resources of larger brands.
Because Shopify fashion merchants are a narrow, niche ICP, the first task was understanding their reality. Before defining anything, I worked to empathise with their perspective: the Shopify workflow they follow day to day, and the wider concerns they carry as brand owners. I then studied competing tools to understand the features and output quality already on offer. Together, these inputs shaped the product decisions.
From there, I owned the end-to-end flow. I worked hands-on alongside the developers, contributing directly in the code as we built it, designed the complete UI/UX, and led outreach to prospective merchants to plan launches and bring more users onto the tool.
Designing the merchant workflow
One of the primary goals was reducing operational complexity for merchants. The generation flow needed to be intuitive enough for first-time users while remaining scalable for merchants managing large catalogs.
This structure let merchants move from product selection to publish-ready content through a guided workflow, rather than navigating multiple disconnected features.
Key product decisions
Story-based generation instead of feature-based generation
Early discussions considered exposing generation capabilities directly through technical controls. But merchants think in terms of outcomes, not AI features. They want catalogue photography, marketing visuals, and videos, not prompt engineering, model selection, or generation parameters.
To simplify decision-making, we introduced story-based workflows where users start by selecting the type of content they want to create (e.g. Catalogue Photography, Video Generation). This approach aligned the product with merchant goals rather than technical functionality, while creating a scalable foundation for introducing new workflows as the product evolved.
Auto-Config for scalable content production
The core promise to Shopify merchants was speed: revamping an entire existing catalog quickly. Manual controls offer flexibility but become unmanageable across hundreds or thousands of products, so we introduced Auto-Config based on theme selection. Merchants set store-level preferences once, and the system reads each product's gender, category, and description to apply suitable generation settings, story-specific logic, and relevant visual configurations, balancing automation and quality while keeping effort minimal.
Challenges & trade-offs
Auto-Config vs. manual selection. Every brand owner has their own requirements and aesthetic, and styling in fashion is deeply subjective: what works for one store can feel wrong for another. Making a single automated solution work across every gender, category, and occasion-type of garment was genuinely hard, and it took many iterations and layers of refinement to reach reliable results. We solved this by expanding the Photogenix library and introducing tag-based matching to auto-configure relevant styling assets.
Story-based, not feature-based. Every merchant runs a different storefront and template, so it isn't possible to satisfy every individual requirement. The challenge was choosing the most valuable capabilities from Photogenix and moulding them to merchant convenience, deciding what truly mattered to these users rather than putting every feature on the plate. That prioritization shaped the whole experience.
Pricing framework. Shopify needed a different pricing model from the core Photogenix product. After researching what would work best for this audience, we set a lower entry rate so merchants can experience the tool's real value with a minimal first recharge and, once they see the results, return for frequent monthly refills.
The outcome: value delivered
For merchants, the value is clear: how quickly they can revamp the imagery already on their Shopify storefront, and how little they have to think to do it.
- Speed. Existing catalog imagery can be refreshed to premium, fashion-standard visuals fast.
- Seamless auto-sync. Outputs sync straight back to the store and reflect on the live website, with no manual re-uploading.
- Scaling without manual effort. Across thousands of products, merchants don't have to pick a model, background, pose, or styling for each one. Auto-Config reads gender, category, and product details and applies relevant, on-standard configurations automatically.
- Native & embedded. The tool lives inside Shopify, so merchants never have to leave their platform.
- Accessible pricing. A researched, lower entry-point framework that lets merchants try the tool with minimal commitment.
The launch also set the groundwork for future improvements in automation quality, onboarding, and generation performance.
What I learned
Moulding an existing product to fit the Shopify case was an interesting thing to learn: stitching the best pieces together rather than forcing the entire app into the Shopify dashboard.
Users care less about the sophistication of the underlying AI and more about how easily a solution fits into their existing workflow. While generation quality mattered, the biggest opportunities often came from reducing friction, simplifying decisions, and helping merchants move from product catalog to publish-ready assets with minimal effort.
The complexity behind simplicity. Building hands-on in the code alongside the developers, I saw first-hand how much engineering it takes to make an experience feel effortless. Every simplification the merchant sees on the surface was a deliberate decision underneath: a back-and-forth of studying what already existed, questioning each step, and stripping it down without losing capability. That gap between a clean interface and the logic holding it up is where I learned the most, both technically and in how I approach UI/UX design.
Encoding subjective aesthetics into logic. The most interesting challenge was translating something as subjective as fashion taste into a system a machine could act on. Rather than leaving styling to chance, I helped shape a tag-matching layer inside the Auto-Config engine, mapping fashion attributes to models, backgrounds, and styling so the right aesthetic choices fired automatically for each product. Turning a designer's intuition into deterministic tag logic taught me how to make creative judgment scalable.
Together, this project deepened my technical fluency and design instinct, and strengthened my understanding of workflow design, product adoption, and the unique challenges of bringing AI capabilities into existing business processes.