Knak / 2026
AI Email Generation
Led design strategy for Knak's flagship AI feature, driving 50% increase in AI adoption across 7,000+ monthly active users. Reduced email creation from days to minutes.
Design Highlights
Overview
I led the design of Knak's flagship AI feature, transforming email campaign creation from a multi-day process into a guided, AI-powered workflow. Powered by Google Gemini, it generates fully editable email drafts complete with structure, copy, and images while maintaining brand consistency.
Challenge
Email marketers traditionally spent days creating email campaigns from scratch—writing copy, sourcing images, building layouts, and ensuring brand consistency. This time-intensive process created bottlenecks in campaign launches and limited the number of campaigns teams could execute. Customers needed a way to radically accelerate email creation while maintaining quality and brand standards.
Approach
I designed an AI workflow that guides users through five steps: prompt submission, design selection, image integration, generation, and refinement. I supported text prompts, PDF briefs, and URLs, then connected the flow to existing brand themes, image libraries, and digital asset managers. The goal was to make AI accessible to marketers without removing meaningful control.
Solution
I designed an interface where users describe their campaign goals and provide supporting materials. The AI analyzes the objective, applies brand themes and guidelines, and generates an editable draft with copy, layout, and imagery. Users can bring in previous email PDFs, reference Brand Voice settings, and pull images from multiple sources before refining the result in Knak's builder.
Impact
The feature reduced email creation from days to minutes, helping marketers launch faster and create more campaign variations. It contributed to a 50% increase in MAU/WAU for AI features and became a key differentiator for Knak in the marketing automation space.
Key Learnings
I learned that successful AI features must balance automation with user control. Customers wanted help generating content, but they also needed to refine the output and understand its limits. Integrating with existing brand themes, image libraries, and DAMs proved as important to adoption as the generation itself.