Challenge
Building FDBK: Lessons from My First AI-Powered Launch
AI coding tools can generate features fast, but they consistently fall short on design and usability. FDBK was my attempt to close that gap, giving designers, vibe coders, and product builders real-time, actionable feedback.
What We’re Tracking
We launched FDBK into beta with the goal of learning where it delivers the most value and where iteration is needed. For now, the impact is measured in conversations with users and the problems they bring to the tool.
Feedback Quality
% of screens improved with actionable feedback before shipping
Workflow Integration
Frequency of use mid-process vs. end-process
Adoption & Reach
Active users per week, with return sessions as the north star
See FDBK in action: From quick hero checks to mobile quirks and complex dashboards, FDBK plugs into your process to deliver structured, instant feedback exactly when you need it. Watch Loom video above.
Use It Anytime You Need a Gut Check
FDBK helps across a range of screens and design moments. Here are a few places it consistently adds value:
1: Screenshot-to-Critique
Upload. Ask. Get feedback. Use it mid-build—no waiting for a review cycle.
2: Structured, Instant Feedback
Every comment is grounded in design systems, UX heuristics, and accessibility standards—no vague “looks off.”
3: Feedback Loop That Fits Your Process
Use FDBK during the design/build process, not just at the end—so you ship with more confidence, not just more speed.
Market Challenge
AI coding tools promise speed, but they consistently fall short on design and usability. They generate the “average” of what they’ve seen—fine for prototypes, but rarely usable in production. Designers, vibe coders, and product managers are left with gaps: cluttered layouts, unclear hierarchy, poor usability. Existing commenting tools don’t solve this either—they assume the process has stopped, when the real need is feedback as you build.
Build Challenge
- Training the system on industry standards (design systems, heuristics, accessibility) so feedback was grounded, not vague.
- Defining a focused MVP instead of overloading it with every possible feature.
- Designing for multiple audiences—designers, product managers, and new builders—without diluting clarity.
- Communicating that FDBK isn’t “just another commenting tool,” but a new type of feedback loop.
- Shipping quickly while building iteration into the process, knowing this was only the first release.
Approach
- Designed FDBK as a chat-first interface where users can upload screens or paste flows and receive real-time critique.
- FDBK’s training drew from design systems, accessibility standards, and UX best practices, giving it the ability to critique hierarchy, layout, clarity, and usability and expanding.
- Focused on making feedback actionable, not vague (e.g., “spacing feels off” → “increase vertical rhythm to improve scanability”).
- Built the tool for builders of all levels, not just trained designers.
Results
- Public beta launched.
- Early users are actively testing, giving feedback on both value and gaps.
- Clear insight: FDBK is not just a UX critique tool, but a partner for people shipping with AI.
Lessons Learned
I learned the importance of scoping tightly for a v1 — smaller, testable features helped me launch faster and validate the concept in real workflows.
I learned how critical iteration is when working with AI tools — the outputs weren’t “done,” but a starting point I had to refine, guide, and adapt.
I learned the value of balancing “maker” and “founder” roles — not just designing, but making technical tradeoffs, writing copy, and thinking about adoption.