AI Based Electricity Demand Forecasting Solution

The Vendor is required to provide comprehensive information on AI-based electricity demand load and supply solutions, focusing on technical approaches, data requirements, integration, scalability, regulatory compliance, and support services to inform procurement strategy.
- Solution should not only forecast load but also integrate electricity supply and demand dynamics to provide a more holistic view of grid conditions.
- This includes accounting for generation variability, distributed energy resources (DER) – both customer-owned and other, demand-side programs such as Conservation Voltage Reduction (CVR) and emerging trends like EV adoption.
- By incorporating both supply and demand factors, the tool should enable more accurate predictions and support decision-making for reliability, resource allocation, and adaptability as grid conditions evolve across different time horizons:
•    Short-term for operational decisions and teal-time balancing
•    Medium-term for maintenance planning and resource scheduling
•    Long-term for capacity planning and investment decisions.
- Discuss any use of advanced AI (like deep learning or reinforcement learning) and how it improves forecast accuracy or insights.
- Specify the time scales the solution can predict. Include deep detail to high-level granularity for short-term forecasts (e.g. intra-day or day-ahead), mid-term (weeks to months ahead), and long-term (multi-year load projections for planning). Indicate how the methodology differs by horizon (if at all) and any limitations on lead times.
- For each data type, note the necessary history (how many past months/years of data) and granularity (e.g. 15-minute intervals, hourly, daily) to train an effective model.
- Provide any benchmarks or statistics on your solution’s forecasting accuracy from similar deployments. For example, share typical error metrics such as Mean Absolute Percentage Error (MAPE) or Root Mean Square Error (RMSE) for day-ahead forecasts.
- At a minimum, this should cover integration with SCADA systems (for real-time operational data), any Advanced Distribution Management System (ADMS) or Energy Management System, and Meter Data Management (for pulling AMI data).

- Questions/Inquires Deadline: January 14, 2026