TaskWise Ai

Turning project chaos into clarity

Comment

Inspiration

Project management wastes hours on manual planning and discovers problems only when deadlines fail. We built AI that thinks ahead,a system that predicts risks and optimize workflows before issues emerge.

What it does

TaskWise transforms project chaos into intelligent execution. Describe your project via voice or text, Gemini 3 based agentic rag system generates professional specs with tasks, deadlines, and dependencies. AI matches tasks to team members using semantic skill analysis, detects duplicate work through vector search, flags quality issues, predicts delivery risks, and generates narrative reports explaining what's actually happening.

How we built it

Four Gemini 3 pillars: (1) Text Generation for project/task creation, (2) Embeddings for skill matching and Qdrant-powered duplicate detection, (3) Analysis for quality scoring and risk evaluation, (4) Narrative Generation for human readable insights.

Challenges we ran into

Prompt engineering for consistent structured output. Tuning embedding thresholds to prevent false skill matches. Balancing Gemini 3's analytical depth with real-time performance through strategic caching.

Accomplishments that we're proud of

Gemini 3 turns "build a mobile app" into 47 actionable tasks instantly. Automatically assigns the right designer based on portfolio embeddings. Warns managers days before critical deadlines become impossible.

What we learned

Embeddings unlock intelligence beyond search. Combining generation with analysis creates proactive systems. The best AI thinks ahead for humans. Executives want narrative insights, not JSON.

What's next for TaskWise AI

Multimodal expansion: extract tasks from meeting recordings and whiteboard photos. Predictive resource allocation before bottlenecks form. Cross-project learning from historical data. Autonomous AI agents that update tasks and resolve blockers independently.

Built With

Share this project:

Updates