Photogenix

AI Edit existed before I got involved. But running client projects through the tool myself, I kept meeting the same gap: the quick, ad-hoc fixes users need after a generation weren't where they should be. I proactively took on a revamp of the post-generation experience. The tech team handed over the API keys, and I owned the rest.

I prototyped UI structures in Google AI Studio and studied the editors users already trust. The sidebar panels of Photoshop and Illustrator inspired the drop-down design that shipped: every capability visible at a glance on the sidebar, each expanding into its controls only when needed. That structure deliberately leaves room for new AI sub-features as they arrive (which is how Crop was added), and it drove consolidation calls too. Features sharing a workflow or API were clubbed, so Retouch merged our earlier Auto and Mask features into one.

More than anything else, this stretch taught me how a tech feature is executed end to end, from API to interface to release.

Photogenix claims to put a brand's entire imagery workflow under one digital umbrella, so the library had to hold everything a traditional shoot would. I studied core studio workflows and drove a rebuild around what styling actually demands: model hairstyles and makeup, a full styling wardrobe across top-wear, bottom-wear and footwear (down to saree), editorial-grade backgrounds, props and finishing elements.

Two additions then made it scale. Fashion brands rarely have one aesthetic. They run multiple mood boards and sub-brands, each with its own models, backgrounds, styling, lighting and poses, so teams can now build their own resource libraries, keeping every sub-brand on its own standard. And a creation card lets users generate any missing asset on the spot and drop it straight into their library (carried into the admin CMS as well), turning a fixed catalog into one that expands with every brand that uses it.

The most recent push came from watching how AI imagery actually gets signed off. Between studios and their clients, and inside the clients' own teams, feedback bounces endlessly across chats and emails, and reworks lose track of which comment belonged to which image.

So we integrated a client-approval flow into Studio, our bulk-production solution: every piece of feedback documented in one place, pinned to the exact image it refers to, and reworkable from right there. I guided the UI structure and pushed for it to work just as well on a phone: reviewing a batch should feel as easy as scrolling through window shopping. A small enhancement on paper, but it makes approvals faster, cleaner and genuinely user-friendly.