Back to Home
Improving a Mobile AI Writing Assistant
My role: Lead Designer
Time: 2023
The case
AI Writer's initial version (v1.0), launched in the early 2023 AI boom, gained initial user traction but faced significant hurdles: poor usability stemming from an unintuitive interface and unclear flows, widespread visual inconsistencies, and a core prompt suggestion feature that lacked practical value, and a suboptimal payment experience contributing to low conversion rates, despite initial adoption. Addressing these challenges became the primary design focus upon taking over the project.
The approach
The strategic approach involved overhauling key user experiences. This included transforming the Home screen into an intuitive template hub to improve feature discovery, replacing ineffective prompts with actionable, data-driven suggestion cards for better guidance, and establishing UI coherence using Atomic Design principles as a foundation, and optimizing the payment flow to increase pay rate. These interventions aimed to significantly enhance the application's usability and overall quality.
My Approach & Key Design Solutions
Taking ownership of AI Writer's Product Design after the initial release, the product team established a refined vision to address v1.0's shortcomings and capitalize on learnings. The key strategic shifts focused on:
Improving Usability
Redesigned onboarding & home screen elements to reduce initial user drop-off. Also optimized the payment flow to address confusion and improve conversion rates.
Revamping Prompt Suggestions
Overhauled the prompt suggestion system by basing content on popular v1.0 prompts and presenting them as intuitive, action-titled cards (e.g., 'Write an email') for better usability.
Establishing UI Consistency
Defined standard colors and built a clear component library based on Atomic Design principles, ensuring visual coherence for AI Writer 2.0 even before a formal Design System was in place.
A friendlier onboarding experience
Variant A - 94% follow-through rate
Variant B - 88% follow-through rate
To optimize the user onboarding experience, we A/B tested two distinct approaches: Variant A (more illustrative and character-driven) against Variant B (a structured step-by-step flow). Variant A proved more effective, yielding a 94% follow-through rate, surpassing Variant B's 88% and validating the impact of a more engaging initial interaction.
Solving the drop-rate problem
User drop-rate
40%
12%
Shifted Concept
From v1.0's input screen to v2.0's template hub, enhancing feature discovery and usability
Effective Prompt Templates
Replaced v1.0's ineffective list with categorized, actionable prompt cards based on user data.
Consistent Interface
Addressed v1.0's visual flaws with standardized components (Atomic Design). Despite minor initial bugs, this improved coherence significantly.
However, certain v1.0 flow choices, like keeping the AI chat with prompt-based flows separated, were intentionally retained for potential future iteration.
The new payment experience
1.0 Payment Screens
2.0 Payment Screens
Post-onboarding Payrate
1%
8%
(approximately)
Upgrade Triggers: Contextual bottom sheets prompted subscription upon hitting credit/feature limits.
Post-Dismissal A/B Test: Tested timed sales pop-up effectiveness after users dismissed the payment screen.
Non-Subscriber Options: Offered 'a la carte' credit purchases and daily rewarded ads (with streak bonus) for free credits.
The new component system and guidelines
Design Components (atomic design)
The new mascot
Addressing v1.0's widespread visual inconsistencies was crucial for the redesign. An initial component library based on Atomic Design principles, along with standardized colors and styles, was implemented for v2.0. While foundational rather than a fully formalized system at this stage, it significantly improved visual coherence and design workflow efficiency. Complementing these structural improvements, a custom-designed brand mascot was introduced to inject personality, making the AI interaction feel more approachable and engaging, particularly for the target non-technical audience.
Conclusion & Reflection
Although the intensive redesign phase for AI Writer was relatively short, this project provided a valuable opportunity to operate as Lead Product Designer and apply a wide range of core methodologies. Key practices included driving improvements through user-centric design principles, analyzing metrics to identify and address critical issues like user drop-off, and validating solutions via A/B testing. It was a rewarding experience in rapidly iterating and enhancing a live product based on both qualitative insights and quantitative results.
Let's create value together
Have a project in mind or want to discuss potential collaborations?
I'd love to hear from you.
Get in touch
hello.hagenbui@gmail.com
bqhungg
I'm currently open to work
Remote / Hybrid / On-site
GMT +7
Fulltime / Freelance
Hanoi, Vietnam