Quillbot is an AI writing platform with 35M+ monthly active users. In 2026, it set a new course: from writing tools to a full creator suite — adding AI Chat and design tools.
Company
Quillbot
Contribution
Design, Research, Strategy
Industry
SaaS, B2C
Year
2026
TL;DR
Quillbot's new AI features were stuck: too expensive to offer unlimited, too limited to satisfy users. The hidden caps were small, and even users ready to pay for more had no way to. I led the design of a credits system that unlocked both sides — every plan includes monthly credits, AI actions spend them, anyone can buy more.
Results: $200K per month in new revenue. 12% of paying users bought extra credits. Shipped as the foundation of Quillbot's new multi-tier pricing.
Context
Quillbot has two parts
Core writing tools
where free users get daily usage caps, like a number of words per day. Or unlimited option for paid users.
New AI tools
chat and design generation — where credits were being introduced.
Credits are simple: every plan includes a monthly amount, and each AI action — a chat message, a generated design — spends some. Run out, and you can upgrade or buy more. This one system now defines what every plan offers, how upgrades happen, and what users see when they reach their limit — for individuals and teams alike. It had to work for 35M+ monthly users, most of them free.
I led the design end to end: from conflicting PRDs to launch. Four teams — product, engineering, marketing, and legal — needed one shared direction. I became the person who set it.
Business goal
Turn "unlimited" pricing into a tiered, sustainable model that grows MRR, protects unit economics on AI features, and holds up legally.
The problem space
Every AI request costs real money
Unlike traditional features, each chat message or image generation has a real cost. With unlimited use, new AI features lost more money the more the product grew.
Invisible caps for new AI features
AI usage was already capped behind the scenes. Users hit invisible walls with no warning, no explanation, and no way out. Even when users wanted to pay more they couldn't — money was being left on the table.
The clock was ticking
New AI features barely had traffic yet — but marketing was about to scale them. Introducing limits after millions rely on something free means taking value away. The system had to come first.
Where I stepped beyond design
Set the scope the organization aligned on
The project started with 2 requirement documents that contradicted each other — down to basics like whether purchased credits expire. I mapped every contradiction and turned it into a prioritized decision list. That list became the launch scope.
Made two limit systems feel like one product
Quillbot's core writing products had their own free-usage caps, and healthy metrics we couldn't risk.
Credits had to sit next to them without users ever asking "which limit did I just hit?" I ran working sessions across product, design, engineering, and marketing to draw the line: what spends credits, what stays as is, and how the product explains the difference. Core metrics held steady through launch.
Built the decision framework, not just the screens
One designer can't review every screen of a system this size. So I wrote the rules instead: voice per balance state, naming for external copy, onboarding logic per segment. The team shipped consistent decisions without me in the room.
The system
I mapped every state and set the rule the system follows: block the AI action, never the writing or manual tools.Users always see their credits — in the navigation, in the account, inside the tools. I mapped every situation: plenty left, running low, ran out, payment failed, credits about to expire — each with its own message and a clear way forward.
One rule everywhere: when credits end, only AI features pause. The writing tools keep working.
A pricing page between 2 eras
The page had to keep selling what Quillbot is known for — its writing products, the conversion core — while introducing credits and preparing a four-tier lineup. I designed the transition: familiar products up front, credits as added value, core conversion untouched.
Three audiences, three introductions.
– New users learn credits as how the product works
– Free users get them as a bonus
– Paying subscribers get 30 days' notice before anything changes.
Same system — three first impressions.
Results
A/B experiment · 5% of traffic · English-speaking markets · all devices
+$200K
revenue / month
north-star target, validated at 5% rollout
2.8%
upgrade conversion
free users at the 100% gate
+12%
top-up conversion
premium users who bought more credits
