-
-
Every score, weight, and weighted contribution is visible so the user can understand exactly why an option ranks first.
-
Changing the Price weight to 60% instantly changes the winner from Laptop A to Laptop C.
-
DecisionSnap makes the decision, criteria, options, and current priorities visible before ranking.
Inspiration
I wanted to build DecisionSnap because most AI recommendations feel like a black box: they tell you what to choose, but not why. I wanted to see exactly why an option wins while keeping the human in control of the final decision.
What it does
DecisionSnap is a small proof of concept where the user can see why an option wins, change what matters, and immediately see how that changes the outcome. The AI is responsible for scoring each option against the chosen criteria, while the final ranking is calculated deterministically from those scores and the user's weights.
How we built it
I built DecisionSnap using Node.js, plain HTML/CSS/JavaScript, and Gemini. The AI is responsible for scoring each option against the chosen criteria, while the final ranking is calculated deterministically from those scores and the user's weights.
Challenges we ran into
The main challenge was handling Gemini errors and making the demo reliable. Another challenge was keeping the criteria weights at exactly 100% while the user adjusts them. I solved these issues by simplifying the UI, adding fallback handling for failed AI calls, and automatically rebalancing the weights. The result is a small proof of concept where the user can see why an option wins, change what matters, and immediately see how that changes the outcome.
Accomplishments that we're proud of
I am proud that DecisionSnap delivers a reliable demo with a simple UI, fallback handling when Gemini fails, weights that always stay at exactly 100%, and a transparent ranking that keeps the human in control of the final decision.
What we learned
While building it, I learned that better prompts and a spec-first approach lead to much more predictable results. I also learned how important it is to actually understand AI-generated code instead of simply accepting it, and to separate what the AI does from the deterministic logic that should remain under my control.
What's next for DecisionSnap
The next step for DecisionSnap is API integration and access to online stores, allowing it to search in real time for the best options currently available based on the user's criteria and priorities.
Log in or sign up for Devpost to join the conversation.