Loading project case study
All work

Knak / 2024-2026

AI Subject Line & Preview Assistant

Helped marketers generate and compare subject-line ideas inside their campaign workflow, reaching 60% monthly account usage and increasing AI usage frequency by 50%.

Read the case study
AI Subject Line & Preview Assistant

Value and impact

More frequent AI use
50%
Accounts using it monthly
60%

01

Overview

I designed Knak's first AI-powered feature: an in-editor assistant that generates tailored subject lines and preview text.

Built on Google Gemini, it analyzes email content, applies Brand Voice rules, and incorporates the user's prompt to create relevant suggestions. The interaction model became the foundation for Knak's later AI features.

  • AI/ML
  • UX Design
  • Knak

Challenge

Email marketers spend significant time crafting compelling subject lines and preview text that drive open rates.

Writing effective subject lines requires balancing creativity with brand consistency, audience relevance, and campaign goals. As Knak's first AI feature, the challenge extended beyond the product itself. There was real hesitation about whether users would trust AI-generated content in their email workflows.

Approach

I placed the assistant directly inside the Knak Email Editor and designed it to offer guidance without boxing users in.

It generates five immediate suggestions, then lets users keep chatting to refine tone, length, and direction. I also added account-level controls so customers could disable the feature while trust was still developing.

04

Solution

Built an AI assistant embedded in the email editor that generates five subject line and preview text suggestions using the context of the email body content, configured Brand Voice attributes and rules, and any additional prompt instructions from the user. The conversational interface allows users to iterate beyond the initial suggestions, refining tone, adjusting length, or exploring different creative directions. The assistant pulls brand context automatically so suggestions are on-brand without requiring users to specify guidelines each time.

01 / 02

Generate suggestions from email content

The assistant uses the email content to generate subject-line and preview-text suggestions, with a clear loading state.

02 / 02

Review and apply paired suggestions

Review subject-line and preview-text pairs and insert both with one action. The editor also shows the active New Subscriber variant.

05

Impact

Marketers can generate alternatives where they already write subject lines, making ideation and A/B testing faster without handing control to an opaque workflow.

The assistant reached 60% monthly active account usage and increased MAU/WAU by 50%, demonstrating that customers would repeatedly use focused AI help inside a familiar task. For Knak, that trust established the interaction and adoption foundation for Brand Voice, translations, and full email generation.

06

Key learnings

Building Knak's first AI feature was a learning experience in designing for trust.

Adding the ability to disable the feature gave customers control, but most chose to keep it on once they experienced the value. The biggest barrier to adoption wasn't AI trust. It was that many customers still rely on external copy creation due to legal approvals and internal policies. Starting with a focused, high-value use case (subject lines) was the right entry point for AI at Knak, establishing patterns and building confidence that enabled more ambitious AI features.

PreviousAI Brand Voice View all projects