AI for Utilities and Energy Suppliers
AI for utilities takes over recurring work in grid connections, market communication and customer service, such as checking connection requests for completeness or drafting replies on dynamic tariffs. NexPatch runs the models on the utility's own premises or in a German data center, aligned with an ISMS under ISO 27019 and with NIS2. Your specialists review and approve.
More cases, same team
The energy transition reaches you as casework: a connection request, a supplier switch, a customer question. In 2025 alone, Germany added 16.4 gigawatts of solar capacity. At the same time, 47% of German energy suppliers report significantly more effort to fill open positions.
AI has arrived in the sector, yet rarely in daily operations. 61% of suppliers that have introduced AI are still in the pilot and experimentation phase. Between pilot and production lie data integration, security and accountability. That is exactly where we come in.
Bundesnetzagentur 01/2026 · BDEW, staffing needs in the energy sector, 2024 · PwC, AI in the energy sector, 04/2026
Regulation and security: what AI at a utility has to meet
Utilities work under stricter security rules than most industries. Three frameworks shape every AI solution, and we plan for them with your ISMS lead before the project starts.
IT security catalogue and ISO 27019
Under Section 11 of the German Energy Industry Act (EnWG), grid and energy plant operators need an ISMS based on ISO/IEC 27001 with the ISO 27019 extensions, including regular audits. An AI solution becomes part of that ISMS. It therefore needs operating documentation, a permission concept and logging from day one.
NIS2
Since December 2025, the German NIS2 implementation act has extended risk management and reporting duties, including for many mid-sized municipal utilities. Outsourcing AI to a public cloud means giving up part of your control, while the responsibility stays with you.
Market communication deadlines
Since June 2025, a supplier switch in Germany has to be completed within 24 hours. Any delay in the inbox or in case assignment becomes a deadline risk.
Run privately: customer and grid data stay in your house
We run language models and data in your own data center or in a German data center, with no dependency on US cloud services. Sign-in runs through your Active Directory, and retrieval only touches documents the user is allowed to see. Every answer comes with a source reference, and every use is logged in an audit-proof way. We use open models, so you can always keep working without us.
Typical flow: from connection request to response
Intake.
A request for a solar system, heat pump or wallbox arrives via portal, email or as a PDF.
Check.
The AI checks completeness and plausibility and spots missing documents.
Follow-up.
The AI requests missing documents from the applicant or installer.
Preparation.
Complete cases are created and prepared in your core system.
Approval.
Your specialists only review exceptions and approve.
Status.
The AI answers status requests automatically, so your team is no longer interrupted.
Integration with your billing and office systems
We connect the AI to the systems you already run. No system change is needed.
Billing and market communication
SAP IS-U, Wilken, Schleupen, SIV, powercloud or Robotron, connected through interfaces instead of copy and paste.
Office and file storage
Microsoft 365, Exchange, SharePoint and file shares serve as knowledge source and as inbox for the AI.
Metering and market data
Meter data, load profiles and market data feed forecasts on our data platform Orpheon.
Use cases at utilities and energy suppliers
Check and answer grid connection requests
Completeness checks, pre-qualification and status updates for solar, heat pump and wallbox requests.
Sort the market communication inbox
Classify, assign and prepare cases from supplier switching and market communication.
Answer questions on dynamic tariffs
Draft replies to tariff and billing questions from your own knowledge base, including the notices required under Section 41a EnWG.
Search the rulebook
An internal knowledge system covering the EnWG, Bundesnetzagentur rulings, the IT security catalogue and your own policies, with a source reference for every answer.
Secure knowledge before it retires
Grid and process knowledge from documents, logs and legacy systems becomes searchable before the colleagues who carry it in their heads leave.
Forecast load and generation
Forecasting and optimization models on your metering and market data, run on our data platform Orpheon.
How a project at a utility starts
Intro call and potential analysis. We review two to three processes that tie up your team the most and pick the one with the clearest benefit.
Pilot in 72 hours. A working first use case on your data. You see results, not slides.
Rollout with your departments. Integration with your systems, roles and approvals, plus user training. We expect 4 to 10 weeks until production.
Production with SLA. We take over monitoring, updates and model maintenance, with 99.9% availability and clear accountability on our side.
For strategy, consulting and training we work with our partner Alic.ai. One engagement, two clearly separated roles, one shared outcome.
Frequently asked questions
In 60 minutes we show which cases in grid connections, market communication and customer service can be automated and what that means for your team.