Decision Maker

Inspiration

I wanted to build something that does more than give quick answers. The idea came from imagining a logical co pilot that actively questions assumptions, challenges weak reasoning, and explores different possible outcomes before settling on a decision. The goal was to make expert level analysis understandable and useful for everyday people, not just specialists.

What it does

Decisiv.AI helps people make tough decisions when the stakes are high. Users can enter complex problems, from major business choices to personal life decisions, and the system breaks them down logically. Instead of vague advice, it produces a clear, structured path forward based on data, trade offs, and likely outcomes.

How I built it

The app uses a clean, glass style interface built with React 19 and Tailwind CSS. It combines multiple AI models, using Gemini 3 Pro for deep analysis, Gemini 2.5 Flash for fast follow up questions, and Veo 3.1 to create visual summaries that make decisions easier to understand.

Challenges I ran into

The biggest challenge was handling rate limits for a system that does a lot of heavy thinking. I solved this by prioritizing the main decision first and loading deeper analysis like audits and simulations only when needed. This kept the system responsive and avoided hitting usage limits.

Accomplishments that I am proud of

Multi Modal Orchestration I successfully combined text, live audio, and real time video input into one smooth experience. Building gapless audio streaming using raw PCM data was a major technical achievement.

Deterministic Accuracy The system produces clear, structured execution plans instead of open ended chat responses. These plans can be synced directly with professional tools like project management software.

The "Thinking" UX Instead of showing a loading spinner, the interface shows how the system is reasoning step by step. This helps users trust the process and understand how conclusions are reached.

Cinematic Visualization Using Veo 3.1, the app generates cinematic video intros for decisions. This turns complex analysis into something engaging and easy to follow.

What I learned

I gained deep experience in coordinating multiple AI systems at once, including live audio, camera input, and structured data outputs. I learned how to work with raw audio streams for real time interaction and how to design prompts that intentionally search for edge cases and unexpected risks.

What's next for Decision Maker

Multi Agent Governance Expanding from a single reasoning system to a group of specialized agents, similar to a virtual boardroom, where different perspectives debate the same decision.

Enterprise Integrations Adding direct connections to tools like SAP, Jira, and Salesforce so decisions made in Decisiv.AI can automatically trigger real world actions.

Historical Back Testing Allowing users to analyze past decisions to spot recurring patterns or biases and improve future recommendations.

Mobile Native "Field Deck" Building a mobile app designed for fast, on site decision making in areas like emergency response and healthcare.

Built With

• gemini-2.5-flash-native-audio-preview-12-2025
• gemini-2.5-flash-preview-tts
• gemini-3-pro-preview
• lucideicons
• mediadevicesapi
• react-19
• recharts
• tailwind
• typescript
• veo-3.1-fast-generate-preview
• webautoapi