brief
↓
Claude → content
↓
Cursor → template
↓
Braze → campaign
Automating the lifecycle messaging pipeline with AI
Designed and built an end-to-end AI pipeline for lifecycle messaging production. Claude generates the content from the brief; Cursor assembles the responsive templates — emails in HTML/CSS + Liquid, in-app messages in HTML/CSS + JavaScript; Braze executes the campaign across both email and in-app messaging. The handoff between steps is structured JSON, which makes the whole flow versionable and reviewable instead of hand-authored every time. The same pipeline feeds the A/B variants — subject-line and CTA tests running at 3M+ sends per arm, analysed in Hex.
<15 min
per campaign — down from 3–8 hours
