The most important terms around artificial intelligence, digital transformation and the energy transition, clearly explained.
73 terms from A to Z
AFIR is the EU regulation on the deployment of alternative fuels infrastructure. Operators of public charging points have had obligations on ad-hoc payment, price transparency and data provision since 2024; from 8 January 2026 ISO 15118-2 is mandatory for new charging points, from 2027 ISO 15118-20 with plug and charge. Utilities operating charging points must adapt back end and processes.
AI systems that autonomously pursue goals, make decisions and execute multi-step tasks without constant human guidance. Unlike traditional AI applications that respond to individual queries, agentic systems actively plan and dynamically adapt their approach. In enterprise practice, this means AI agents that take over entire workflows, not just individual steps.
AgNeS is the Federal Network Agency's ruling procedure for a new electricity grid charge system that replaces the StromNEV from 2029. It provides for time-variable charges, capacity prices and feed-in charges. The consultation ended in September 2026, the framework ruling follows by end of 2026, follow-up rulings in 2027. Grid operators have to adapt billing, metering concepts and tariffs.
An autonomous software system that perceives its environment, makes decisions and takes actions to achieve specific goals. AI agents differ from simple chatbots through their ability to use tools, create plans and independently execute multi-step tasks. In enterprises, they are increasingly deployed for research, data analysis and process automation.
An application programming interface through which software systems exchange data and functions in a structured way. APIs are the connective tissue of modern IT, and in AI the primary path to models: applications call language models, embeddings or agents via APIs. Well-designed, documented and secured APIs determine integrability, scaling and vendor independence.
Hypothetical AI systems that can understand and solve any intellectual task that a human can accomplish, including reasoning, learning and creative problem-solving across domains. AGI does not exist today; current systems like LLMs are specialised (narrow) AI. The debate around AGI is central to long-term AI strategies and regulation.
Germany's Federal Office for Information Security is the national cybersecurity authority. It defines standards (such as IT-Grundschutz), issues threat warnings and publishes guidance, increasingly on the secure use of AI in enterprises and public administration. For regulated sectors, BSI requirements are a de-facto compliance benchmark.
Beyond Visual Line of Sight, drone operation outside the pilot's direct field of view. BVLOS flights can cover whole line corridors in one go and are the prerequisite for cost-effective drone inspection of power grids. In the EU they fall into the specific category of drone regulation 2019/947 and require an operational authorisation from the national aviation authority.
Clean Core is the SAP principle of keeping the standard core of an S/4HANA system free of custom developments and building extensions decoupled on the Business Technology Platform. For energy utilities this means assessing, cleaning up or encapsulating decades-old IS-U custom logic so that upgrades, MaKo releases and AI functions remain possible without special effort.
CLS management bundles the processes by which an active external market participant sends control commands to control boxes, heat pumps, wallboxes and storage via the CLS channel (Controllable Local Systems) of the smart meter gateway. It covers roles, keys, communication profiles and evidence. For control under §14a EnWG it is the operating basis of the grid operator or a commissioned service provider.
The systematic design and optimisation of the context available to an AI model during processing. This includes system prompts, example data, document selection and instruction design. Context Engineering goes beyond Prompt Engineering because it shapes the entire information architecture around the model, not just the individual input.
The control box is the BSI-certified device that receives control commands via the smart meter gateway and passes them to controllable consumption devices or generation plants, for example dimming a heat pump to 4.2 kW under §14a EnWG or curtailing a PV system under the solar peak law. Grid operators and metering point operators have to organise ordering, installation and operation at scale.
A customer installation is an energy installation behind the grid connection that supplies several final consumers without being regulated as a grid, for example in apartment buildings or business parks. After the CJEU judgment C-293/23 and the Federal Court of Justice decision of May 2025 the term is interpreted narrowly; since 2026 grandfathering applies and new constellations must be organised as a grid or as tenant electricity.
Control over where data is stored and processed and which jurisdiction governs it. For European organisations it means being able to run critical data and AI workloads independently of non-European providers and legal regimes. It is a core driver of sovereign-AI and EU cloud initiatives.
The default metering point operator is usually the distribution system operator and carries the rollout obligations under MsbG with quotas of 20, 50 and 95 percent by 2025, 2028 and 2030. Competitive metering point operators take over metering points at the customer's request. Since 1 July 2026 standardised metering contracts (BK6-24-125) apply to default operators.
AI agents deployed as full team members with defined roles and responsibilities within organisations. The innobu Digital Colleagues framework gives AI agents concrete job titles such as Architect, Strategist or Analyst and integrates them into existing governance structures. The approach solves the most common problem in AI projects: the technology works, but nobody agrees on what it should actually do.
A digital secondary substation is a medium-to-low-voltage substation with sensors, telecontrol and communication that delivers low-voltage measurements to the control centre and allows remote switching. Large grid operators are retrofitting tens of thousands of substations by 2026; they are the core of grid state estimation for §14a and low-voltage redispatch.
The deep redesign of an organisation's business models, processes and culture using digital technologies, more than digitising individual workflows. Successful transformation ties technology, data and workforce enablement to clear business goals. AI is currently the strongest driver and test of this change.
A virtual representation of a physical object, process or system that is synchronised with real-time data. In the energy sector, digital twins are used to model grid infrastructure, simulate loads and plan maintenance proactively. The value comes from connecting sensor technology, data platforms and decision support.
A digital twin of a district heating network is a hydraulic-thermal model of the network that is continuously reconciled with measurements from substations and generators. Utilities use it to lower the supply temperature, to integrate decentralised feed-in and large heat pumps, and as the data basis for heat planning and the decarbonisation roadmap.
An energy data space is a federated infrastructure under Gaia-X principles in which market actors share energy data sovereignly without handing it over centrally. At EU level the Common European Energy Data Space (CEEDS) is emerging; in Germany the lighthouse project energy data-X with four transmission system operators ran until October 2026. For utilities it is about interfaces, data quality and data governance.
Energy sharing under §42c EnWG allows energy communities to jointly use and settle locally generated electricity within a grid area. Since 1 June 2026 grid operators must enable the settlement; the Federal Network Agency platform under §20b EnWG makes mass operation realistic only from 2027.
The complete description of an organisation's structures: business processes, IT systems, data flows and technology infrastructure. Enterprise Architecture creates transparency about how individual systems interact and provides the decision basis for investments in technology and transformation. Without EA, every large programme becomes a blind flight.
The European AI regulation (Artificial Intelligence Act) that establishes a risk-based regulatory framework for the deployment of AI in the EU since 2024. Systems are classified into risk categories, from minimal to unacceptable. High-risk systems are subject to strict requirements for transparency, data quality and human oversight. For enterprises, the EU AI Act means concrete compliance obligations that must be implemented now.
EU Regulation 2019/947, which governs the operation of civil drones across Europe. It splits flights into the open, specific and certified categories. Commercial BVLOS inspections of power grids fall into the specific category and need an operational authorisation based on SORA, a standard scenario or a PDRA.
The gas network transformation plan is a gas distribution operator's plan for converting its network to hydrogen, biomethane or decommissioning by 2045, aligned with municipal heat planning. The H2vorOrt and DVGW guide sets the methodology; the EnWG gas package 2026, implementing the EU gas directive, creates the legal frame for decommissioning and repurposing.
The smart meter gateway administrator operates and manages the smart meter gateways on behalf of the metering point operator: configuration, certificates, firmware, tariff use cases and secure communication under BSI TR-03109. A GWA must be certified by the BSI; from 1 January 2027 new and re-certifications follow TR-03109-6 version 2.0.
The EU's General Data Protection Regulation has governed the processing of personal data since 2018. It requires a legal basis, purpose limitation, data minimisation and data-subject rights, with fines of up to 4 percent of global annual turnover. For AI projects it is the central framework: training data, prompts and model outputs containing personal data must be GDPR-compliant, forcing anonymisation, data-processing agreements and impact assessments.
AI systems that can create new content such as text, images, code or music, rather than just analysing or classifying existing data. The best-known examples include ChatGPT, Claude and Midjourney. Generative AI is typically based on large pre-trained models (foundation models) and has fundamentally changed the discussion around AI in enterprise contexts since 2023.
A grid connection portal is the digital platform through which connection users and installers apply for grid connections and generation plants, submit documents and track status. Under §8(7) EEG grid operators have had to offer such a portal since 2025; the grid package provides for capacity maps and reservations from 135 kW from 2027. It is the most visible digitalisation process between grid operator and customer.
Grid state estimation is the calculation of the current state of a low-voltage grid, meaning voltages, currents and loading, from measurements at a few points plus a grid model. The VDE FNN guidance recommends measurements from at least 20 percent of connection users, or transformer feeder plus 10 percent, at one-minute resolution. It is the data basis for grid-oriented control from 2029.
Grid-oriented control is the form of control under §14a EnWG that becomes binding from 2029: the distribution system operator intervenes only in an actually measured congestion, based on grid state data from the low-voltage grid. Until end of 2028 the transition period with preventive control by fixed rules applies. Prerequisites are grid state estimation, control boxes, CLS management and a low-voltage control system.
The phenomenon where AI models generate plausible-sounding but factually incorrect or entirely fabricated information. Hallucinations are not a bug but a systemic characteristic of generative models that operate on probabilities. For enterprise use, this means every AI output needs verification mechanisms. RAG, grounding and human oversight are the most important countermeasures.
The Heat Planning Act obliges municipalities to carry out municipal heat planning: cities above 100,000 inhabitants by 30 June 2026, all others by 30 June 2028. The plan designates areas for heat networks, hydrogen networks and decentralised supply. Under §32 operators of existing heat networks must submit a decarbonisation roadmap by end of 2026. Utilities provide the data basis and network planning.
HEDWIG is the Federal Network Agency's data portal under §111g EnWG, to which grid operators and other market actors must deliver pseudonymised energy market data so that authorities and research can evaluate the energy market. The consultation on data scope and pseudonymisation ran until May 2026; the delivery obligation requires new interfaces and data protection processes.
A home energy management system controls heat pump, wallbox, storage and PV system in a building and receives the grid operator's §14a EnWG limits via EEBUS and the CLS channel of the smart meter gateway. Instead of dimming each device individually, the grid operator only tells the HEMS the permitted total power.
The Federal Network Agency's IT security catalogue under §5c EnWG obliges grid operators to run an ISMS under ISO 27001 for grid control technology and, since the 2026 version, to operate attack detection systems in OT networks. Implementation must be evidenced; it affects control technology, telecontrol and the connection of digital secondary substations.
The KRITIS umbrella act has regulated the physical and organisational protection of critical facilities since 17 March 2026, complementing digital cybersecurity under NIS2. Operators must register with BBK and BSI, carry out risk analyses and submit resilience plans. The original registration deadline was dropped; the ordinance sets the dates.
Large neural networks trained on extensive text data that can understand and generate natural language. LLMs such as GPT-4, Claude and Gemini form the foundation of the current generative AI wave. They work with tokens, use the Transformer architecture and are optimised for specific applications through fine-tuning and Reinforcement Learning from Human Feedback (RLHF).
A term coined by Simon Willison for the dangerous combination of three capabilities in one AI agent: access to private data, exposure to untrusted content, and the ability to communicate externally. When all three meet, a prompt injection can exfiltrate data unnoticed. The takeaway: design agents so the three never coexist.
The Luftfahrt-Bundesamt is the German authority that grants operational authorisations for drone flights in the specific category. For new BVLOS applications the LBA has assessed them on the basis of the SORA 2.5 methodology since 1 January 2026; processing usually takes several weeks to months.
The MaBiS hub is the central platform for electricity balancing settlement defined by the Federal Network Agency, replacing today's decentralised aggregation at distribution system operators. The introduction is split into two blocks: ruling on block 1 (meter data processing) in 2026, go-live in the second half of 2028. Grid operators and suppliers have to adapt interfaces and master data.
A subfield of artificial intelligence where algorithms learn from data and recognise patterns without being explicitly programmed. The three main categories are supervised learning (with labelled data), unsupervised learning (pattern recognition in unstructured data) and reinforcement learning (learning through reward). Machine learning is the foundation of virtually all modern AI applications.
Market communication is the standardised electronic data exchange between the market roles of the German energy sector, such as grid operator, supplier and metering point operator, for processes like supplier switching, balancing and meter data transfer. The Federal Network Agency defines the processes (GPKE, GeLi Gas, MaBiS, WiM) and formats (EDIFACT, in future API) and publishes two binding releases a year, on 1 April and 1 October.
MiSpeL is the Federal Network Agency's ruling on the market integration of storage and charging points. It regulates how bidirectional storage and charging points are balanced and which grid charges apply, and removes the double burden of grid charges. The ruling is published on 1 October 2026; automation in market communication follows from 2027.
An open standard that enables AI models to access external data sources and tools in a structured way. MCP was initiated by Anthropic and works like a universal adapter: instead of building a custom integration for each data source, MCP provides a standardised interface. For enterprises, this reduces integration effort and improves interoperability between AI systems.
An architecture in which multiple AI agents collaborate to solve complex tasks. Each agent has specialised capabilities and a defined role; coordination happens via protocols or an orchestration agent. Multi-agent systems enable the decomposition of complex problems into manageable sub-tasks and are the foundation for scalable AI operating models.
A subfield of AI concerned with the processing, understanding and generation of human language. NLP encompasses tasks such as text classification, sentiment analysis, named entity recognition, machine translation and question answering. Modern LLMs have revolutionised NLP: tasks that required specialised models just a few years ago are now solved by generic language models as a side effect.
NEST refers to the Federal Network Agency's rulings of December 2025 on the new incentive regulation replacing the ARegV. The fifth regulatory period starts in 2028 for gas and 2029 for electricity, followed by three-year periods. New are quality elements on digitalisation and energy transition capability that may carry revenue consequences in future.
The EU's NIS2 Directive tightens cybersecurity obligations for operators of essential and important entities, including energy, health, digital infrastructure and public administration. It mandates risk management, incident reporting within 24 hours and personal liability for management, pulling far more organisations into scope than its predecessor.
Software whose source code is open to inspect, use and modify. In AI the spectrum ranges from open weights to fully open training data and recipes. Open-source models offer control, auditability and sovereignty, but shift responsibility for operation, security and compliance in-house.
A middle path defined by EASA between a standard scenario and a full individual SORA. A PDRA describes a typical operation with risk already assessed, so operators use the ready-made template instead of producing the complete risk analysis themselves.
The art and science of formulating input instructions (prompts) for AI models to achieve optimal results. This includes techniques such as chain-of-thought prompting, role instructions, format specifications and few-shot examples. Good prompt engineering can drastically improve the quality of AI outputs without needing to modify the model itself.
An attack vector where malicious instructions are injected into the input of an AI model to manipulate its behaviour. In indirect prompt injection, instructions are hidden in data sources that the model processes, such as websites or documents. Prompt injection is one of the biggest security challenges for AI systems in enterprise deployment.
Quality regulation is the part of incentive regulation that assesses grid operators via indicators such as SAIDI (interruption duration) and, in future, degree of digitalisation and energy transition capability. In July 2026 the Federal Network Agency published a map of Germany with these indicators for the first time, without revenue consequences yet; under NEST they may become bonuses and penalties.
An architecture that enriches AI models with external knowledge by retrieving relevant documents before generating a response. Instead of relying solely on training knowledge, the system searches a knowledge base and inserts relevant information into the context. RAG reduces hallucinations, keeps answers current and enables the use of proprietary enterprise data without retraining the model.
Redispatch is the grid operator's intervention in the output of generation plants to avoid grid congestion. Since Redispatch 2.0 (2021) plants from 100 kW and controllable small plants are included, with data exchange via Connect+. Redispatch 3.0 extends the procedure to market-based flexibility and small plants in the low-voltage grid (VDE SPEC 90032).
A framework for developing and deploying AI systems with consideration for ethical principles such as fairness, transparency, privacy and non-discrimination. Responsible AI goes beyond compliance: it requires proactive measures such as bias testing, explainability of decisions and clear accountability structures. In the EU context, it is increasingly becoming mandatory through the AI Act.
The metric that relates the financial return of an investment to its cost. For AI projects ROI is often hard to pin down: productivity, quality and speed gains pay off indirectly while pilot costs land immediately. Credible ROI measurement needs a clear baseline, defined use cases and honest accounting for operating, governance and change costs.
SAP IS-U is the industry-specific billing and market communication solution many German utilities run on SAP ECC. Mainstream maintenance for ECC ends on 31 December 2027; extended maintenance until 2030 costs a surcharge. The successor is S/4HANA Utilities, usually with a clean-core approach; migrations take 18 to 36 months and depend on MaKo releases and rollout.
Section 14a of the German Energy Industry Act (Energiewirtschaftsgesetz) regulates the grid-oriented control of controllable consumption devices such as heat pumps, wallboxes and battery storage. Grid operators can temporarily dim power to 4.2 kW during bottlenecks, but must grant reduced grid fees in return. Implementation requires Smart Meter Gateways, new control boxes and significant adaptations to IT systems and processes at grid operators.
The central communication unit of an intelligent metering system (iMSys) that enables secure data exchange between electricity meters, market participants and grid operators. Smart Meter Gateways must be BSI-certified and form the technical foundation for the German smart meter rollout. They are the prerequisite for time-variable tariffs, Section 14a control and the integration of decentralised energy generation.
Specific Operations Risk Assessment, the EU methodology for assessing the risk of drone flights in the specific category. SORA evaluates ground and air risk and derives the safety measures required. Since 1 January 2026 new BVLOS applications in Germany must use version SORA 2.5.
A predefined operating scenario in the EU drone regulation (STS-01, STS-02) that sets approved standard conditions for drone flights in the specific category. Operators who stay within a standard scenario can declare the operation to the aviation authority instead of running an individual SORA risk analysis.
Tariff use cases are the operating modes of a smart meter gateway defined in the BSI protection profile; they determine which measurements are transmitted to whom at which resolution, for example TAF 7 for meter reading profiles in 15-minute values for dynamic tariffs and balancing, or TAF 6 for readings on demand. New gateway generations support more TAFs in parallel.
The smallest processing unit in language models, a word, word fragment or special character into which text is broken down. An English word typically consists of 1 to 2 tokens. Tokens determine both the input and output length of a model and are the basis for cost calculation when using APIs. A model's context length (e.g. 200,000 tokens) determines how much information can be processed simultaneously.
A neural network architecture introduced in 2017 that is based on the self-attention mechanism and forms the foundation of virtually all modern language models. Transformers can capture relationships between distant words in a text and can be trained in parallel. GPT, Claude, Gemini and Llama are all based on this architecture. The name of the paper: "Attention Is All You Need".
The European regulatory framework for digital airspace management of drones. U-space provides services such as identification, geo-awareness and traffic information so that many drones can fly safely and automatically in the same airspace, for example along power line corridors.
Dependence on a single vendor whose products, formats or interfaces make switching technically or economically expensive. In AI, lock-in arises through proprietary models, APIs, embeddings and tooling. Countermeasures include open standards, abstraction layers, exit strategies and multi-vendor architectures that keep models and platforms interchangeable.
A software development approach where developers describe their intentions in natural language and AI generates the code. The term was coined in 2025 by Andrej Karpathy. Vibe Coding drastically lowers the entry barrier for software development and enables domain experts without deep programming skills to build functional prototypes and applications. Quality assurance remains the critical human task.
A security vulnerability exploited before the vendor can ship a patch, defenders have zero days of warning. Zero-day exploits are especially dangerous because neither signatures nor updates protect against them. For AI systems the attack surface additionally shifts onto models, prompts and agent tooling.
Techniques where AI models can solve tasks without (zero-shot) or with only a few (few-shot) examples. In zero-shot, the model receives only a task description; in few-shot, a few example pairs (input/output) are provided. This capability is a key advantage of large language models over traditional machine learning, which requires large labelled datasets.
LFW24 is the process, binding since 2025, under which an electricity supplier switch must be completed within 24 hours. The switch runs star-shaped via the market location ID at the grid operator instead of bilaterally between suppliers. Extension to gas and hydrogen is planned; the date is open.