🚀 Inspiration We were inspired by how overwhelming and outdated compliance management is for energy companies. Regulations are dense, constantly changing, and still managed with spreadsheets. We wanted to create an AI-driven system that transforms weeks of manual review into seconds of intelligent automation.

⚙️ What it does

EnergyGrid.AI Compliance Copilot automatically analyzes regulatory PDFs, extracts key obligations, generates actionable compliance tasks, and produces executive reports—all in under 30 seconds using AWS Bedrock and Nova models.

🧠 How we built it

We built a three-agent system (Analyzer, Planner, Reporter) powered by Amazon Bedrock AgentCore.

Analyzer uses Claude 3 Sonnet for NLP extraction Planner uses Nova Pro for task reasoning Reporter compiles structured reports Frontend: Streamlit Backend: AWS Lambda, DynamoDB, S3, and Cognito for security.

🧩 Challenges we ran into

Configuring multi-agent coordination using Bedrock AgentCore, optimizing context memory, and integrating real-time updates between agents were the toughest parts. Ensuring compliance-grade accuracy under 30 seconds was another major challenge.

🏆 Accomplishments we’re proud of

We built a fully functional AWS-native agentic app capable of real-time regulatory reasoning. Our system reduced compliance analysis from 3 weeks to 30 seconds—a massive leap for the industry.

📚 What we learned

We mastered Bedrock AgentCore, Strands SDK, and Kiro, learning how to orchestrate autonomous agents with reasoning, planning, and reporting pipelines inside AWS’s ecosystem.

🌍 What’s next for EnergyGrid.AI

We plan to launch enterprise pilots with energy utilities, expand to multi-language regulatory datasets, and release an open-source compliance RAG agent framework for developers worldwide.

Built With

Share this project:

Updates