WDAIDT — Why Did AI Do That?

Inspiration

AI assistants can produce unexpected answers by overlooking instructions, making assumptions, or missing important context. We built WDAIDT to help users understand why these responses happen and how to improve their prompts.

What it does

WDAIDT is a Chrome extension that analyzes ChatGPT prompts and responses to identify overlooked instructions, missing constraints, assumptions, and trade-offs. It generates structured diagnostic reports and suggests improved follow-up prompts to help users get more relevant and accurate responses.

How we built it

We developed WDAIDT as a Chrome extension that analyzes conversations using prompt-analysis logic and generates structured explanations. We implemented manual and automatic analysis modes, configurable mock mode, and diagnostic reports with actionable follow-up prompts.

Challenges we ran into

One major challenge was identifying meaningful explanations without claiming access to an AI model's hidden reasoning. We also worked on extracting conversation context, structuring diagnostic results, and making the analysis useful without overwhelming users.

Accomplishments that we're proud of

We built a working Chrome extension concept that turns unexpected AI responses into opportunities for learning. We're proud of its structured diagnostic reports, follow-up prompt suggestions, and flexible analysis modes.

What we learned

We learned how to build browser extensions, analyze prompt-response interactions, design structured AI diagnostics, and distinguish observable response patterns from assumptions about a model's internal reasoning.

What's next for WDAIDT — Why Did AI Do That?

We plan to introduce conversation history, prompt version tracking, diagnostic comparisons, support for more AI platforms, and personalized recommendations. Our goal is to help users identify recurring prompt issues and build better prompting habits over time.