What I Learned Building a One-Photo AI Photoshoot Workflow

Short summary

A product builder shares lessons from creating GenBlink, an AI photoshoot workflow that turns a single portrait into curated photo packs. Key decisions include separating curated art direction from user-facing prompts, implementing transactional credit reservations, and treating batch history as a core product surface with per-image status and partial failure recovery. The core insight is that the contract around the model—input boundaries, progress visibility, honest failure handling—matters more than the model itself.

• •Separate curated creative direction from user prompts to keep AI image generation predictable
• •Implement transactional credits: reserve before generation, charge only on success
• •Treat batch history and per-image status as a core product surface, not an afterthought

Generated with AI, which can make mistakes.