GenAI UX
AI Product Design
B2B Platforms
Human-AI Collaboration
Project Overview
PromptPal.AI is a modular, AI-powered prompt-building tool that helps professionals like UX writers, designers, marketers, and content teams create structured, reusable prompts for generative AI.
This project was driven by a goal to make AI tools more intentional, transparent, and usable in real-world workflows bridging the gap between creative control and machine intelligence.
Role:
Team:
Timeline:
From messy prompts to modular clarity
a tool that gives user a full control over
how they write with AI.
Context & Background
Problem
What was the problem?
Generative AI tools are powerful, but they aren’t designed for how professionals like UX writers, designers, and marketers work. People often have to start from scratch, reuse prompts inconsistently, or copy them from scattered sources. This creates confusion, extra effort, and makes it hard to get consistent, reliable results especially when clarity and control are important.
These gaps create friction for users who need consistency, reuse, and clarity when designing or writing with AI.
Solution
I envisioned PromptPal.AI as a tool that brings structure, transparency, and modularity to the creation of prompts. Instead of a single input box, PromptPal allows users to:
Create prompts using blocks (Role, Task, Tone, etc.)
Add AI Memory to persist and reuse context across sessions
View structured AI output with explainability features
Save and insert reusable Prompt Packs and blocks
This transforms AI prompting from a black box into a creative, editable interface.
I led the entire UX process from research and experience mapping to wireframes, UI design, and testing. I used my understanding of GenAI to design features that give users more control and clarity. I also observed how others work with AI tools and tested early designs with peers, using their feedback to improve the overall experience.
01
Understanding the Problem
I began by identifying pain points users faced with traditional GenAI tools: scattered prompt workflows, lack of consistency, and low trust in AI outputs. I reviewed existing prompt management behaviors, studied expert workflows (UX writers, marketers, researchers), and mapped where cognitive friction occurred in prompt reuse and editing.
02
User Interviews
Conducted 6 user interviews with content designers, product managers, and marketers from my LinkedIn network.
I asked targeted questions and assigned a short task to gather insights into their workflows and challenges.
03
Heuristic Evaluation
Conducted a heuristic evaluation of prompt-based workflows in enterprise GenAI tools(ChatGPT, Claude, Jasper) to uncover usability gaps and cognitive friction. Assessed interfaces based on established usability principles, such as visibility of system status, consistency, user control, and error prevention.
Focused on how professionals like designers, marketers, and analysts create, reuse, and manage prompts in daily workflows. Insights revealed issues around prompt discoverability, lack of feedback, and limited modularity, leading to recommendations for clearer affordances, reusable prompt libraries, and better system transparency.
04
Experience Mapping
Using insights from user interviews and competitive analysis, I mapped out the full journey of crafting, editing, testing, and saving prompts. This helped define key moments where user agency and explainability were most critical, and where memory or reuse would reduce friction.
AI UX Principles for GenAI
To guide design, I applied core AI UX principles:
Explainability
User Autonomy
Let users control what’s remembered, reused, or revised.
Reusability
Help users build prompt libraries they trust and iterate on.
About
System Thinking
Key design decisions included:
Prompt Blocks instead of a single input field - to give clarity and modularity.
Save to Memory toggle – to give users control over what gets stored.
Output Tags & Explanations – to add transparency to GenAI results.
Agent Suggestions & Draft Modifiers – to reduce manual prompt rewriting.
I designed PromptPal as a modular system: each “block” (prompt, memory, modifier) acts like a reusable, editable unit. This structure allowed for flexible workflows while keeping the interface minimal and focused.
Modular Block System
AI Memory
Introduced a Memory Panel where users could store reusable inputs like:
Audience = “First-time founders”
Platform = “Mobile app”
These memory items could be turned on/off depending on the needs of the session. This flexibility helped prevent irrelevant context from influencing the prompt.
Explainability and Feedback
Each AI output came with:
“Why this?” anchors explaining what influenced the text
Block-by-block highlighting to match output lines
Thumbs-up/down for refining results
In testing, users reported this helped them better understand how to get repeatable, high-quality output.
Save to Library
Why: Most users reused parts of the prompts repeatedly.
Impact: Improved speed and consistency.
Use from saved blocks
Added a feature to insert from saved blocks, allowing users to quickly reuse previously created prompt elements like Tone or Role. This supported fast, consistent composition.
Challenges & Iterations
Early versions overwhelmed users with an excessive number of visible controls. Users didn't understand what “Blocks” meant or how “Memory” worked.
What I Changed:
1. Simplified onboarding
2. Renamed terminology (e.g., “Block” → “Prompt Block”)
3. Added tooltips and mini-tutorials
Built “Suggested Blocks” and templates for beginners
