Unexpected demand peaks
A shift overlap, a heat wave, or equipment starting at once can set a monthly peak in fifteen minutes — and demand charges may follow it for the whole billing period.
Energy forecasting & planning · In development
Plan tomorrow’s energy demand today.
LoadPilot is being built to help facility and energy teams anticipate electricity demand, spot likely peaks, and weigh tariff and battery options before the day begins — with every plan approved by an operator.
Next-day demand · Site A
Illustrative
Forecast peak—
Likely window—
Charge window—
Sample forecast
An illustrative planning view
A preview of how a next-day plan could look: forecast versus actual demand, the anticipated peak, and suggested battery charging windows. Switch between facility types to see different load shapes.
Illustrative data — not a real facility
Sample factory · Tuesday
Forecast peak——
Forecast daily energy—± confidence range shown in chart
Forecast error so far (MAPE)—Actuals through 14:00
Suggested charging—Pending operator approval
Hover or focus the chart and use ← → keys to inspect half-hour values.
Suggested battery plan Draft · requires approval
Illustrative 500 kWh battery, 250 kW max rateView chart data as a table
| Time | Forecast | Low (10th pct) | High (90th pct) | Actual | Battery plan |
|---|
All values are generated for demonstration. Real forecasts will depend on each site’s meter history, operating schedule, weather, and data quality.
The problem
Energy decisions are often made after the fact
Many facilities see their demand clearly only when the bill arrives. That makes it hard to act on the things that drive cost and risk.
Limited planning visibility
Meter data, production schedules, and weather usually live in separate systems, so teams plan tomorrow using yesterday’s spreadsheets and rules of thumb.
Hard-to-coordinate sources
Grid supply, time-of-use tariffs, on-site generation, and battery capacity each have their own constraints. Weighing them together by hand is slow and error-prone.
Planned features
What LoadPilot is designed to do
Capabilities in the current development roadmap. Scope may change based on feedback from pilot facilities.
Electricity demand forecasting
Half-hourly, next-day and week-ahead demand forecasts per site, built from meter history, operating calendars, and weather.
Peak-demand alerts
Early notice when forecasts suggest a site may approach a threshold you set, with the likely time window and contributing factors.
Tariff scenario comparisons
Estimate how a forecast day would price under different tariff structures or load-shift plans, side by side, with the assumptions shown.
Battery scheduling recommendations
Suggested charge and discharge windows that respect your battery’s capacity and rate limits — presented as a draft for an operator to review.
Forecast confidence ranges
Every forecast comes with a range, not just a single line, so teams can see how certain a prediction is before acting on it.
Site-level reporting
Per-site summaries of forecast accuracy, peaks, and approved plans, exportable for facility, finance, and sustainability teams.
How it works
From meter data to an approved plan
-
1
Import meter & operating data
Upload interval meter exports, production or occupancy schedules, and equipment details such as battery size and limits.
-
2
Generate forecasts
Models combine site history with weather to produce demand forecasts and confidence ranges for each site.
-
3
Compare scenarios
Test tariff options, shifted schedules, or battery strategies against the forecast and review the estimated trade-offs.
-
4
Approve an operating plan
An operator reviews, edits, and approves the plan. Nothing is actioned on site without that sign-off.
Planned AWS architecture
Built on managed AWS services
The planned data pipeline, subject to change as development progresses.
-
Store
Amazon S3
Durable storage for raw meter interval data, weather data, and uploaded operating schedules.
-
Prepare
AWS Glue
Cleans, validates, and aligns meter, weather, and calendar data into model-ready datasets.
-
Forecast
Amazon SageMaker AI
Trains and runs demand forecasting models that output predictions with confidence ranges.
-
Schedule
AWS Lambda
Runs scheduled processing and generates draft peak alerts and battery recommendations for review.
Planned Draft recommendations are returned to the LoadPilot app for operator review. There is no direct connection to site equipment in the planned design.
Operating approach
Decision support, not autopilot
LoadPilot is designed to inform the people who run your facility — not to replace their judgement or operate equipment on its own.
-
Operators approve every plan
Recommendations are drafts. A named person reviews and approves them, and remains in control of all equipment.
-
Results depend on local conditions
What a plan can achieve depends on your tariff structure, the equipment you have — such as battery capacity or flexible loads — and your operating constraints.
-
Data quality matters
Forecasts are only as reliable as the meter history and schedules behind them. Gaps and uncertainty are shown, not hidden.
-
No guaranteed savings
We show estimates with their assumptions so teams can make informed decisions. We don’t promise specific outcomes.
About
Where LoadPilot is today
LoadPilot is an early-stage product focused on practical energy planning for industrial and commercial sites. We are building with input from facility and energy managers and are looking for a small number of pilot facilities to shape the first release.
Development status
- Complete Product concept & planning workflows
- In progress Forecasting prototype & data pipeline
- Next Pilot programme with early facilities
- Later General availability
LoadPilot has no production customers yet. Timelines are indicative.
FAQ
Common questions
Is LoadPilot available now?
Not yet. LoadPilot is in development. We’re inviting a limited number of facilities to join an early pilot and help shape the product.
Does LoadPilot control my equipment?
No. LoadPilot produces forecasts and draft recommendations. Operators decide whether and how to act, and all equipment stays under your existing controls.
Will LoadPilot reduce my energy bill?
We can’t guarantee savings. Any benefit depends on your tariffs, equipment, flexibility, and data quality. LoadPilot is designed to make trade-offs visible so your team can make better-informed decisions.
What data would a pilot need?
Typically 12 months or more of interval meter data (15- or 30-minute), basic operating schedules, and — if relevant — details of batteries or on-site generation and your current tariff.
How accurate will the forecasts be?
Accuracy varies by site and data quality. Every forecast includes a confidence range, and site reports will track how forecasts compared with actual demand over time.
Where will data be stored?
The planned architecture stores data on AWS (Amazon S3). Specific region, retention, and access terms would be agreed with each pilot facility.
Pilot programme
Request early access
Tell us about your facility. If it’s a good fit for the pilot, we’ll get in touch to discuss data, scope, and timelines. There’s no commitment.
- Factories, warehouses & commercial buildings
- Interval meter data available
- A facility or energy lead to review plans