• 

  Account Type Choosing Interface

• 

  Upload Electricity Bill Interface

## Inspiration

Malaysian SMEs face rising electricity costs but only receive static bills that explain what they owe, not why costs increased or what action to take. Meanwhile, community solar operators struggle to identify suitable businesses for solar participation and load shifting. We built EnergiKita to bridge this gap by turning electricity bills into actionable energy intelligence.

What it does

EnergiKita FlexOS transforms electricity bills and basic business information into a verified Energy Flexibility Passport.

The platform:

• Extracts and validates bill data using AI
• Analyses energy trends and potential waste
• Identifies flexible energy activities
• Estimates savings opportunities
• Generates personalised recommendations
• Matches SMEs with community solar programmes
• Optimises energy schedules across businesses, solar, batteries, and EV charging
• Audits every recommendation before approval

How we built it

Built with:

• Frontend: Next.js, React, TypeScript, TailwindCSS
• Backend: FastAPI, PostgreSQL
• AI: Multi-agent architecture for bill extraction, validation, analysis, passport generation, matching, auditing, and explanations
• Optimisation: OR-Tools for constraint-based scheduling
• Forecasting: Demand and solar prediction models

Core workflow:

Bill → Energy Profile → Flexibility Passport → Community Matching → Optimisation → Audited Impact

Challenges we ran into

• Preventing AI from making unsupported energy claims
• Generating useful insights from limited monthly bill data
• Balancing technical sophistication with a simple user experience
• Ensuring recommendations remain operationally realistic through auditing and validation

Accomplishments that we're proud of

• Created the Energy Flexibility Passport concept
• Built a true multi-agent workflow with audit and retry loops
• Combined AI, forecasting, and optimisation into one platform
• Produced explainable, confidence-scored recommendations
• Demonstrated community-level energy coordination using low-cost data inputs

What we learned

• AI alone is not enough; deterministic validation and optimisation are equally important
• Transparency and confidence scoring build user trust
• Energy recommendations must respect real business constraints
• Multi-agent systems work best when each agent has a clear responsibility

What's next for M & N_EnergiKita

Near Term

• Pilot with SMEs and community solar operators
• Validate real-world savings
• Improve forecasting accuracy

Mid Term

• Integrate smart meters and interval data
• Enable continuous optimisation and monitoring

Long Term

• Become Malaysia's SME energy intelligence layer
• Coordinate large-scale community energy ecosystems
• Support regulated surplus-energy and demand-flexibility programmes

Vision: Make sophisticated energy intelligence accessible to every SME without requiring expensive hardware from day one.

Built With

• c
• fastapi
• grafilabapi
• next.js
• postgresql
• python
• react
• tailwind
• typescript