NRG.01
Renewables Permitting
Wind, solar and storage permitting support: EIA chapters cross-checked, consultation responses coded and conditions tracked, with specialists signing every environmental conclusion.
Open pageIndustry
Energy AI,
with an engineer on shift.
Reads the grid. Never drives it. Renewables permits, load and generation forecasts and customer operations for utilities and energy companies. The models read from your operational systems and never write to grid control; an engineer or operator decides.
Last reviewed 30 October 2026 · Engineering guidance, not legal or regulatory advice.
In context
AI in energy means using models to read permitting evidence, forecast demand and renewable output, and handle customer contact for suppliers and network operators. We build it outside the control loop: models inform, and a qualified person acts.
Where it breaks today
01
Projects wait on paperwork
A wind or solar project can spend years in permitting and grid connection queues. Much of that time is document work: environmental statements, consultee responses, conditions and the evidence that each one has been met.
02
Variable supply, harder forecasts
Embedded solar, heat pumps and electric vehicles have changed the shape of demand. Forecasts built for a system of large thermal plants miss more often, and the cost lands in balancing and imbalance charges.
03
Customer contact at peak stress
Price changes, estimated bills and smart meter faults arrive in waves. Some of those callers are in financial difficulty or depend on electricity for medical equipment, and missing that is a regulatory and human failure.
OUR APPROACH
AI for utilities, outside the control loop
Energy is critical infrastructure, and the EU AI Act treats it that way: AI used as a safety component in the management and operation of electricity, gas or heating supply is listed as high-risk in Annex III. NIS2 places energy among its sectors of high criticality. Both point to the same design choice, which is ours too. The systems we build read from SCADA historians, metering and billing systems; they do not write set points, switch plant or dispatch generation.
Inside that boundary there is a lot of useful work. Language models are good at permitting documents and customer contact. Forecasting is mostly a job for well-tested statistical and machine-learning models, with a language model at most explaining why a forecast moved. We say which is which, and we do not dress a gradient-boosted forecast up as generative AI.
How AI can help renewable energy specifically: by shortening the document side of development and by forecasting variable output with honest uncertainty, so operators can plan around it. Whether a given system falls under the high-risk regime is a determination for your compliance function; we document the architecture so it can be made.
HOW IT DEPLOYS
PHASE 01
Map data and boundaries
Which systems hold the data (AVEVA PI, SAP IS-U, Oracle CC&B, Kraken, GIS), which are operational technology, and where the read-only boundary sits.
PHASE 02
Build against a baseline
Every forecast is backtested against your current method and a naive baseline; every document workflow is measured on your own past files.
PHASE 03
Operate with accountability
Outputs go to the engineer, trader, planner or agent who owns the decision, with sources and uncertainty shown, and drift is monitored after launch.
What we build
01
Renewables Permitting
Environmental statements, consultee responses and permit conditions read, cross-checked and tracked, so a planning team spends its time on judgement rather than collation.
SEE RENEWABLES PERMITTING →02
Load and Generation Forecasting
Probabilistic demand, solar and wind forecasts from weather, calendar and metering data, backtested against your current method and delivered with ranges rather than single numbers.
SEE LOAD FORECASTING →03
Utility Customer Operations
Billing queries, meter reads, moves and payment plans resolved against the billing system, with vulnerability and financial difficulty detected and routed to trained staff.
SEE CUSTOMER OPERATIONS →04
Asset Records and Maintenance
Inspection reports, work orders and asset histories made searchable and summarised for planners, alongside the predictive maintenance models your engineers already trust.
SEE PREDICTIVE MAINTENANCE →Controls we implement
READ-ONLY TO OPERATIONAL SYSTEMS
NO WRITES TO GRID CONTROL
NO TRAINING ON YOUR DATA
SMART METER DATA UNDER GDPR CONTROLS
FORECASTS BACKTESTED BEFORE USE
AUDIT LOG ON EVERY OUTPUT
These are engineering controls we build into the architecture, not certifications. QQuantum.ai holds no security certifications, and the audits and attestations your sector requires remain yours to run — see our trust centre.
Energy and utilities use cases
-
-
NRG.02
Load & Generation Forecasting
Probabilistic demand, solar and wind forecasts backtested against your current method and a naive baseline, delivered as ranges to the operators and traders who decide.
Open page -
NRG.03
Utility Customer Operations
Utility customer service on the real account in SAP IS-U, Oracle CC&B or Kraken, screening every contact for vulnerability and financial difficulty before resolving anything.
Open page
Questions
Questions about Energy & Utilities.
How is AI used in the energy sector?
In four main places: forecasting demand and renewable output, maintaining assets from sensor and inspection data, handling customer contact and billing, and the document work of developing and permitting projects. The control of the grid itself stays with operators and certified control systems. Most of the practical value today is in forecasting quality and in the time saved on documents and customer queries.
How can AI help renewable energy?
Two ways matter most. It shortens project development by reading and cross-checking the environmental and consultation documents that permitting depends on. And it forecasts variable solar and wind output with honest uncertainty ranges, so traders and system operators can plan around it. Both need a person accountable: a planning lead for the application, an operator for dispatch.
Is AI in critical energy infrastructure high-risk under the EU AI Act?
AI used as a safety component in the management and operation of electricity, gas or heating supply is listed as high-risk in Annex III of the EU AI Act, which brings risk management, logging and human oversight obligations. Systems that only draft documents or answer customers usually fall elsewhere. The classification of a specific system is your compliance determination; we design and document so it can be made clearly.
Do your systems connect to SCADA?
Read-only, and usually not directly. We read operational data from a historian such as AVEVA PI or a data platform downstream of it, through the segmentation your operational technology security programme defines. Nothing we build writes set points or switches plant. That boundary is a design decision, not a configuration option.
What about the energy AI itself uses?
It is a fair question for an energy company. We size models to the task, prefer smaller and open-weight models where they perform well enough, and can run inference on your own infrastructure or in a region with a low-carbon grid. Forecasting models in particular are computationally light compared with large language models.
