WithPromptly
AI has been everywhere for the past couple of years. We use it for everything — fixing grammar, brainstorming ideas, even writing stuff we don’t feel like writing ourselves. But somewhere along the way, we forgot how to work with it properly. As a designer, I felt this pain even more. The design process has so many steps, and we often just throw vague prompts like “make this better” and hope for magic. Spoiler: it doesn’t work.
"It was how we were talking to it."
I was attending an AI-focused event in Bangalore. Tools were being demoed, but what stood out was the frustration. Everyone was asking AI for help and getting vague, unusable responses. The common reaction was to blame the tool. But watching closely, I realized the real issue wasn't the AI—it was our lack of structure. That moment became the starting point for WithPromptly.
The Problem
Forgotten Clarity
Heavy reliance on AI has weakened our ability to structure thoughts. Vague inputs = vague outputs.
Chat ≠ Search
Most users treat AI like Google. But AI needs context, structure, and intent, not just keywords.
Process Skipping
Designers often jump straight to screens, skipping foundational steps. This causes endless iteration later.
What I Built
WithPromptly is a prompt library built specifically for designers. Each prompt forces structured thinking, mirrors real design process steps, and produces consistent AI outputs.
Instead of asking AI to do the work, Promptly helps users think better while working with AI.
Built For
- Design Students
- Early-career UI/UX Designers
- Non-designers learning Design Thinking
- AI power users struggling with output
The Challenges
1. Writing prompts that actually work
The hardest part. I tested roughly 100 prompts across 5 different AI tools. I studied prompt engineering patterns and research papers to find the balance between "too rigid" (robotic) and "too loose" (vague).
2. No-Code Accessibility
I don't come from a dev background, so I built this as a no-code system intentionally. Constraints: Easy access, no login friction, simple copy-paste flow.
3. Avoiding "Prompt Spam"
Quantity over quality is a risk. I consciously kept the library small and intentional, focusing only on prompts that solved real design steps and reduced iteration time.