Designing GhostWriter and the Labs Environment for a Platform Used by Over a Million Brands

The insight that drove both designs

AI content tools fail when they require users to leave their context to use them. If a marketer is building an email in Tailwind and needs to generate copy, they shouldn’t have to open a separate tool, generate text, and paste it back in. That was the workflow most platforms had in 2022 — and it was slow and disjointed.

The brief for both GhostWriter and Labs was the same underneath: AI should show up where the user already is, at the moment they need it, rather than requiring them to navigate to a separate destination.

GhostWriter — AI writing inside the editor

Email body copy generation

GhostWriter sits inside the email editor. When a user is about to write body copy, the AI option is already there — no mode switch, no separate screen. The user provides a prompt or lets GhostWriter infer from campaign context already in the system (subject line, audience, campaign type), gets a draft back, and edits from there.

The draft appears in place and is immediately editable. The experience was designed to feel like a smart collaborator alongside the writing process, not a separate tool you consult and copy-paste from.

Design challenge: generated text is variable in length and quality. The design had to cover the full range — a great draft that needs minor tweaks, a draft that needs significant reworking, and a draft that misses entirely and needs a new prompt. The interface handles all three gracefully without making the simple case unnecessarily complex.

Subject line generator

Subject lines are high-stakes — they determine open rates and benefit from testing and iteration. The subject line generator was designed to produce multiple options simultaneously rather than one draft at a time: direct, curiosity-based, emoji-led, question format. Users can see several approaches and choose or iterate from the one closest to what they want.

Labs — a modular home for AI capabilities

The problem Labs was solving

The product team was adding AI capabilities faster than the core product could absorb them. Without a designated place, AI features would spread inconsistently across the product, each with its own interaction pattern and its own discoverability problem. Labs was designed to be that designated place.

Lab Cards

The core structure of Labs is the Lab Card — one card per AI capability. Each card gives users enough context to understand what the tool does and lets them launch into using it. The card structure means new capabilities slot in as new cards without requiring a navigation change or a redesign.

Cards at launch included:

• GhostWriter for email body copy — the same tool available in the editor, accessible as a standalone starting point
• Subject line generator — for generating subject lines outside a specific email
• Article wizard — a multi-step tool for generating longer-form content: blog posts, newsletter articles, content requiring more structure than a single prompt
• Social content tools — caption and post generation for social media scheduling

Browser extension

Labs also ships as a browser extension — giving users access to the AI toolkit without being inside Tailwind at all. Someone drafting a social post on LinkedIn or a product description in their e-commerce platform can use the Labs tools from the extension without switching context. This extended the reach of the AI capabilities beyond Tailwind’s own surfaces without requiring users to change their workflow.

What launched

GhostWriter and Labs shipped to Tailwind’s full user base. Users who had been leaving the platform to generate content elsewhere had a reason to stay inside the product. The Lab Cards structure gave the product team a framework to ship new AI capabilities incrementally — which proved prescient as AI content tools continued to develop rapidly after launch.