BIE Energy-Flow Agent

Autonomous agent for BesmaInfoEnergy: uses Gemini & MongoDB MCP to optimize gas/electricity distribution, predict failures, and act on the grid in real-time via Google Cloud.

Comment

Inspiration

As a Computer Systems Engineer with deep roots in the Algerian electricity and gas distribution sector, I have witnessed firsthand the limitations of traditional, passive monitoring systems. With BesmaInfoEnergy, my inspiration was to bridge the gap between legacy infrastructure and Modern AI. The goal is to move beyond simple dashboards and create a "conscious" grid where energy flows are optimized autonomously, reducing waste and predicting critical failures before they happen.

What it does

BIE Energy-Flow Agent is an autonomous intelligent agent designed for the utility sector.

  • Real-Time Monitoring: It tracks electricity voltage and gas pressure levels across distribution points using MongoDB Atlas.
  • Predictive Intelligence: It identifies anomalies such as voltage drops or gas pressure surges that could indicate equipment failure.
  • Agentic Action: Unlike a standard chatbot, this agent uses the Model Context Protocol (MCP) to "act." It can automatically generate maintenance tickets, alert field technicians, or suggest load-balancing strategies to protect the infrastructure.

How we built it

The project architecture integrates high-performance cloud and data technologies:

  • Google Cloud Agent Builder: Used to orchestrate the agent’s reasoning and workflow capabilities.
  • Gemini 3: The core brain providing advanced reasoning to analyze complex energy patterns.
  • MongoDB Atlas: Our central data nerve center, storing consumption logs and asset details.
  • Model Context Protocol (MCP): The critical bridge that allows Gemini to securely query and write to our MongoDB database in real-time.
  • Python & VS Code: For building our distribution simulation scripts and integration layers.

Challenges we ran into

The primary challenge was implementing the MCP server to ensure seamless communication between Gemini's reasoning engine and the structured data in MongoDB. Harmonizing heterogeneous data from two different utility types—Electricity and Gas—required a robust data schema to ensure the agent could make safety-critical decisions accurately.

Accomplishments that we're proud of

We are proud to have transformed a "static" database into a dynamic environment where an AI agent can read, reason, and act within seconds. Successfully integrating decades of industry expertise into a cutting-edge Agentic AI framework is a major milestone for BesmaInfoEnergy.

What we learned

This hackathon deepened our mastery of the "Agentic" paradigm. We learned that the future of utilities isn't just about better data visualization, but about agents capable of using specialized tools to solve real-world industrial problems autonomously.

What's next for BIE Energy-Flow Agent

Our next step is to integrate real IoT sensor hardware to move from simulation to a live pilot within a local distribution network. We also plan to expand the agent's capabilities to include solar energy management, offering a comprehensive, AI-driven green energy optimization suite for the Algerian market.

Built With

  • besmainfoenergy
  • gemini-3
  • google-cloud-agent-builder
  • model-context-protocol-(mcp)
  • mongodb-atlas
  • n8n
  • node.js
  • python
  • vs-code
Share this project:

Updates