Brand Playbook
AI tooling reached Shopify faster than any shared practice for using it. This is the productized answer: a structured library of tested prompts and playbooks covering brand voice, product copy, and merchandising — packaged as a tool rather than a blog post.
The problem
Merchants got powerful AI tooling and no method. The useful prompts live in scattered threads and screenshots — untested, uncredited, and impossible to apply consistently across a brand. The gap isn’t access to AI, it’s knowing what to ask and keeping the answers on-brand.
The approach
Treated it as a product, not content. Built a taxonomy first (brand voice, product copy, merchandising, customer replies), then wrote and tested every prompt against real Shopify stores through Claude’s MCP connection, so the outputs were checked against live catalogue data rather than guessed at. Designed the library so a merchant can find the right prompt, understand why it works, and adapt it to their own voice instead of pasting someone else’s.
The outcome
A packaged playbook that holds a brand’s voice steady across everything AI touches — the strategy, brand, and design layers shipped as one tool.
A few pieces of the build
Outcomes worth measuring
- ProductizedTool, not a doc
- 4Playbook categories
- StrategyBrand + design
Have a project like this?
Send the link or the screenshot you keep meaning to share — I'll come back with where the seams are and what's worth fixing first.