THE ENTERPRISE AI FIELD GUIDE
Understand the language.Connect the concepts. Find a concept Learning level All learning levels Essential Practitioner Advanced
All concepts 188 concepts AI Foundations · ● Essential Artificial intelligence Systems performing tasks associated with human intelligence, such as prediction, perception or language processing.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Machine learning Foundations · ● Essential ML Learning patterns from data rather than specifying every rule manually.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Deep learning Foundations · ◇ Practitioner DL Machine learning using neural networks with multiple layers.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Neural network Foundations · ◇ Practitioner A model made of connected computational units whose weights are learned from data.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Generative AI Foundations · ● Essential GenAI Models that generate new text, images, audio, video or other outputs from learned patterns.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Predictive AI Foundations · ● Essential Models that estimate outcomes, classes or numerical values from input data.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Foundation model Foundations · ● Essential A broadly trained model that can be adapted to multiple downstream tasks.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. LLM Foundations · ● Essential Large language model A large model trained to process and generate language; some also handle other modalities.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. SLM Foundations · ◇ Practitioner Small language model A comparatively smaller language model; there is no universal parameter cutoff.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. NLP Foundations · ◇ Practitioner Natural language processing Methods for analysing, understanding and generating human language.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. AGI Foundations · △ Advanced Artificial general intelligence A proposed form of broadly capable general intelligence; definitions and measurement remain contested.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. ASI Foundations · △ Advanced Artificial superintelligence A hypothetical intelligence exceeding human capabilities across a broad range of tasks.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Model vs application Foundations · ● Essential The model supplies learned capabilities; an application combines it with interfaces, data, tools and controls.Enterprise context: An enterprise team chooses an appropriate AI approach before choosing a vendor. Transformer Inside the model · ◇ Practitioner A neural-network architecture using attention to process relationships in sequences.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Attention Inside the model · △ Advanced A mechanism that weights which parts of an input are relevant to a representation.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Parameters Inside the model · ◇ Practitioner Learned numerical values that shape model behaviour; parameter count alone does not determine quality.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Weights Inside the model · ◇ Practitioner Learned parameters governing computations inside a model.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Token Inside the model · ● Essential A unit used to encode input or output, such as part of a word; token is not synonymous with word.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Tokenizer Inside the model · ◇ Practitioner The mechanism that converts content to model tokens and back.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Embedding Inside the model · ◇ Practitioner A numerical representation that captures useful features or relationships of content.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Context window Inside the model · ● Essential The maximum context a model can process in an invocation; it is not persistent memory.Enterprise context: A platform team compares capability, deployment size and licensing constraints. MoE Inside the model · △ Advanced Mixture of experts An architecture routing inputs to subsets of specialist model components.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Dense model Inside the model · △ Advanced A model architecture where the main parameter blocks participate for each token, unlike sparse expert routing.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Open weights Inside the model · ● Essential Model weights are available under a licence; source code, training data and unrestricted use may not be.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Open-source AI Inside the model · ◇ Practitioner AI released under terms and with materials meeting an applicable open-source definition; inspect the actual licence.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Model card Inside the model · ● Essential Documentation of a model’s intended uses, evaluations, limitations and other relevant characteristics.Enterprise context: A platform team compares capability, deployment size and licensing constraints. Pretraining Training and adaptation · ◇ Practitioner Initial broad training that builds a model’s general capabilities.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Training vs inference Training and adaptation · ● Essential Training learns or updates parameters; inference uses a trained model to produce results.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Fine-tuning Training and adaptation · ● Essential Additional training that changes model weights for a desired task, behaviour or domain.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. SFT Training and adaptation · ◇ Practitioner Supervised fine-tuning Training on examples of desired input-output behaviour.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. RL Training and adaptation · △ Advanced Reinforcement learning Learning a policy through reward signals from actions or generated outputs.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. RLHF Training and adaptation · ◇ Practitioner Reinforcement learning from human feedback Using human preference feedback to help guide model training, often through a learned reward model.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. RLAIF Training and adaptation · △ Advanced Reinforcement learning from AI feedback Using AI-generated feedback as a reward or preference signal.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. DPO Training and adaptation · △ Advanced Direct preference optimization Training directly on preferred and dispreferred responses without a separate RL optimisation loop.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. PEFT Training and adaptation · △ Advanced Parameter-efficient fine-tuning Adapting a model by training a limited subset or additional parameters.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. LoRA Training and adaptation · △ Advanced Low-rank adaptation A PEFT technique that learns small low-rank weight updates.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Distillation Training and adaptation · ◇ Practitioner Training a student model to reproduce useful behaviour from a teacher model.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Quantization Training and adaptation · ◇ Practitioner Representing weights or activations at lower numerical precision to reduce resource requirements.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Synthetic data Training and adaptation · ◇ Practitioner Artificially generated training or evaluation data; quality and representativeness need checks.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Overfitting Training and adaptation · ◇ Practitioner Learning training-specific patterns so strongly that performance on new data suffers.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Data contamination Training and adaptation · ◇ Practitioner Training data overlaps evaluation content, potentially inflating reported performance.Enterprise context: A team adapts output style while keeping current company knowledge in a retrieval system. Prompt Generation and reasoning · ● Essential Instructions and input provided to a model.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. System instructions Generation and reasoning · ● Essential High-level directions used by an application to guide model behaviour; precedence depends on the system.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Completion Generation and reasoning · ◇ Practitioner Generated output responding to an input, historically often a continuation of text.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Temperature Generation and reasoning · ◇ Practitioner A sampling setting that changes output variability; low values do not guarantee correctness.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Top-p Generation and reasoning · △ Advanced Nucleus sampling Sampling from a set of candidate tokens covering a specified cumulative probability.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Hallucination Generation and reasoning · ● Essential An output that invents or misstates information, sometimes with convincing detail.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Structured output Generation and reasoning · ● Essential Output constrained to a defined format or schema; format compliance does not ensure factual accuracy.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Reasoning model Generation and reasoning · ● Essential A model designed or trained for extended problem-solving; it can still make reasoning errors.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Chain of thought Generation and reasoning · ◇ Practitioner CoT Intermediate reasoning expressed as steps; visible steps are not a guaranteed faithful account of internal computation.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Test-time compute Generation and reasoning · △ Advanced Inference-time compute Additional computation during use, such as more reasoning or candidate evaluation.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Self-consistency Generation and reasoning · △ Advanced Comparing multiple generated reasoning paths or answers to select an agreed result.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Multimodal Generation and reasoning · ● Essential Processing or generating more than one modality, such as text, images and audio.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. VLM Generation and reasoning · ◇ Practitioner Vision-language model A model combining visual information with language understanding or generation.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. OCR Generation and reasoning · ◇ Practitioner Optical character recognition Extracting machine-readable text from images or scanned documents.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. ASR / TTS Generation and reasoning · ◇ Practitioner Automatic speech recognition / text-to-speech Converting speech into text / converting text into spoken audio.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Diffusion model Generation and reasoning · △ Advanced A generative model learning to reverse a noise process, widely used for image generation.Enterprise context: An assistant returns validated JSON for a service ticket instead of free-form prose. Prompt engineering Prompts and context · ● Essential Designing and testing instructions and examples to improve model responses.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Context engineering Prompts and context · ● Essential Designing what information reaches a model: instructions, retrieved content, tools, history and state.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Zero-shot Prompts and context · ◇ Practitioner Requesting a task without showing worked examples in the prompt.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Few-shot Prompts and context · ◇ Practitioner Providing a small number of examples to guide the model during an invocation.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Prompt template Prompts and context · ◇ Practitioner A reusable instruction structure with fields populated for a task.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Prompt chaining Prompts and context · ◇ Practitioner Passing outputs through a sequence of model calls with different responsibilities.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Context compaction Prompts and context · ◇ Practitioner Summarising or restructuring context to fit space limits while retaining important information.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Context rot Prompts and context · △ Advanced An informal term for degraded use of relevant information in long or cluttered context.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Prompt caching Prompts and context · ◇ Practitioner Reusing computation for repeated prompt prefixes where supported; distinct from caching final answers.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. In-context learning Prompts and context · ◇ Practitioner Adapting response behaviour from examples in context without updating model weights.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Instruction hierarchy Prompts and context · ◇ Practitioner Rules for resolving instruction priority; untrusted retrieved text should not override trusted instructions.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. Context budget Prompts and context · ◇ Practitioner Allocation of available tokens across instructions, sources, history, tools and output.Enterprise context: A workplace assistant receives the right policy excerpts, user permissions and tool descriptions. RAG Knowledge and retrieval · ● Essential Retrieval-augmented generation Retrieving relevant external information and providing it to a model to support generation; it does not retrain the model.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Grounding Knowledge and retrieval · ● Essential Anchoring an output in supplied sources or evidence; source quality still matters.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Chunking Knowledge and retrieval · ◇ Practitioner Splitting content into retrievable sections with enough meaning and context.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Vector database Knowledge and retrieval · ◇ Practitioner A store supporting search over vector representations, often used with embeddings.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Semantic search Knowledge and retrieval · ● Essential Retrieval based on meaning or learned similarity rather than exact word matches alone.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Keyword search Knowledge and retrieval · ◇ Practitioner Retrieval based on words and lexical matching, useful for exact names and identifiers.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Hybrid search Knowledge and retrieval · ◇ Practitioner Combining lexical and semantic retrieval signals.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Reranking Knowledge and retrieval · ◇ Practitioner Reordering retrieved candidates with an additional relevance model or scoring method.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Metadata filtering Knowledge and retrieval · ◇ Practitioner Restricting retrieval using fields such as date, region, document type or access rules.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Knowledge graph Knowledge and retrieval · ◇ Practitioner A representation of entities and their relationships.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. GraphRAG Knowledge and retrieval · ◇ Practitioner RAG approaches using graph structures or graph-derived summaries to improve retrieval and synthesis.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Agentic RAG Knowledge and retrieval · ◇ Practitioner Retrieval where an agent selects searches, sources or follow-up steps dynamically.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Provenance Knowledge and retrieval · ● Essential The origin and processing history of data or an answer’s supporting evidence.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Permission-aware retrieval Knowledge and retrieval · ● Essential Enforcing user access rights while retrieving content, including indexes and downstream results.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. Retrieval evaluation Knowledge and retrieval · ◇ Practitioner Measuring whether retrieval finds relevant evidence before evaluating the generated answer.Enterprise context: A policy assistant retrieves only documents the employee is authorised to access. AI agent Agents and orchestration · ● Essential A system using a model to choose actions and tools toward a goal within configured boundaries.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Agentic AI Agents and orchestration · ● Essential AI systems with goal-directed action and varying autonomy; industry usage is not uniform.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Copilot Agents and orchestration · ● Essential A product pattern where AI assists a person; actual autonomy varies by implementation.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Workflow Agents and orchestration · ● Essential A predefined sequence or graph of tasks, possibly containing model calls and conditional branches.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Autonomy Agents and orchestration · ● Essential The degree of action a system can take without intervention; define it per task and risk.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Planning Agents and orchestration · ◇ Practitioner Forming and revising steps to achieve a goal.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Tool calling Agents and orchestration · ● Essential Function calling A model requests an operation using a defined tool interface; application code executes the operation.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Orchestration Agents and orchestration · ◇ Practitioner Coordinating models, tools, state and task execution.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. ReAct Agents and orchestration · △ Advanced Reasoning and acting A pattern interleaving model reasoning with actions and observations.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Multi-agent system Agents and orchestration · ◇ Practitioner Multiple agents collaborating or specialising; more agents do not automatically improve performance.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Handoff Agents and orchestration · ◇ Practitioner Transferring task responsibility and relevant context to another agent or person.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Agent memory Agents and orchestration · ● Essential Information retained for later use, such as history or preferences; requires scope and privacy controls.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Agent state Agents and orchestration · ◇ Practitioner Current task data, intermediate results and execution status.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Agent harness Agents and orchestration · ◇ Practitioner The surrounding runtime for tools, context, execution, safeguards and task lifecycle.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Agent skills Agents and orchestration · ◇ Practitioner Reusable instructions and resources for specialised tasks; packaging varies across systems.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Human-in-the-loop Agents and orchestration · ● Essential HITL A person participates in decisions, reviews or approvals at defined points.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. Stop conditions Agents and orchestration · ◇ Practitioner Rules ending execution, such as success, timeout, cost budget or maximum steps.Enterprise context: A service agent investigates an issue, proposes a change and waits for approval before executing it. API Connections and actions · ● Essential Application programming interface A defined interface through which software requests data or operations.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. SDK Connections and actions · ◇ Practitioner Software development kit Libraries and utilities for building against a platform or interface.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. MCP Connections and actions · ● Essential Model Context Protocol An open protocol connecting AI applications to external tools, resources and prompts; it does not itself grant trust or permissions.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. MCP host / client / server Connections and actions · ◇ Practitioner The application hosting AI / its connection component / the provider exposing capabilities.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. MCP resources / tools Connections and actions · ◇ Practitioner Resources expose context or data; tools expose callable operations.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. A2A Connections and actions · ◇ Practitioner Agent2Agent protocol An open protocol supporting communication and interoperability between agents.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Agent discovery Connections and actions · ◇ Practitioner Finding agents and learning their capabilities and interaction requirements.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Connector Connections and actions · ● Essential An integration linking an application with an external system; not every connector uses MCP.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Computer use Connections and actions · ● Essential AI interaction with graphical interfaces, often using screenshots, clicks and typing.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Sandbox Connections and actions · ● Essential An isolated execution environment limiting impact; isolation strength depends on implementation.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. OAuth Connections and actions · ◇ Practitioner A framework for delegated authorisation, often used to give scoped access without sharing passwords.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Least privilege Connections and actions · ● Essential Granting only the access needed for a particular task and duration.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Agent identity Connections and actions · ● Essential An identifiable principal or credential context for agent access, audit and accountability.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Idempotency Connections and actions · ◇ Practitioner Designing repeat requests to avoid duplicate effects, such as issuing the same refund twice.Enterprise context: An agent reads a ticket through a scoped API and cannot change unrelated systems. Responsible AI Responsible AI and security · ● Essential Practices for developing and operating AI with attention to fairness, safety, privacy, transparency and accountability.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. AI governance Responsible AI and security · ● Essential Decision rights, policies, ownership and oversight across the AI lifecycle.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. AI ethics Responsible AI and security · ● Essential Consideration of moral consequences, values and impacts on people and society.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Trustworthy AI Responsible AI and security · ● Essential AI assessed against qualities such as reliability, safety, security and accountability; not a blanket certification.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Fairness / bias Responsible AI and security · ● Essential Evaluating unjust differences or systematic skew; fairness criteria can conflict and depend on context.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Explainability Responsible AI and security · ● Essential Making outputs or behaviour understandable to relevant audiences; generated explanations require scrutiny.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Transparency Responsible AI and security · ● Essential Disclosing relevant information about AI use, capabilities, data and limitations.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Accountability Responsible AI and security · ● Essential Assigning responsibility and mechanisms for review, challenge and remediation.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. AI safety Responsible AI and security · ● Essential Reducing harmful behaviour and consequences during development and use.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Alignment Responsible AI and security · ◇ Practitioner Seeking model behaviour consistent with intended goals, instructions and values.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Guardrails Responsible AI and security · ● Essential Controls that constrain, detect or respond to undesirable inputs, actions and outputs.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Prompt injection Responsible AI and security · ● Essential Malicious or untrusted content attempting to redirect model behaviour, including through retrieved documents.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Jailbreak Responsible AI and security · ◇ Practitioner An attempt to bypass a model’s behavioural restrictions.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Data exfiltration Responsible AI and security · ● Essential Unauthorised transfer of information out of its permitted environment.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. PII Responsible AI and security · ● Essential Personally identifiable information Information that identifies or can be linked to a person; applicable legal definitions vary.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. DLP Responsible AI and security · ● Essential Data loss prevention Controls to detect and restrict inappropriate disclosure or movement of sensitive data.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Data residency Responsible AI and security · ● Essential Where data is stored or processed; check each service and data flow.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Red teaming Responsible AI and security · ◇ Practitioner Adversarial testing to discover weaknesses and harmful behaviours.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. AI impact assessment Responsible AI and security · ● Essential A structured review of intended use, affected people, risks and mitigation measures.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Shadow AI Responsible AI and security · ● Essential AI used outside approved organisational visibility or controls.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. NIST AI RMF Responsible AI and security · ◇ Practitioner AI Risk Management Framework A voluntary framework organised around Govern, Map, Measure and Manage.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Content credentials Responsible AI and security · ◇ Practitioner Provenance information about media creation and edits; provenance does not prove every claim is true.Enterprise context: A governance review examines risk, access, affected people and accountability before rollout. Evals Evaluation and assurance · ● Essential Evaluations Repeatable tests of whether an AI system meets specific quality, safety and task requirements.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Benchmark Evaluation and assurance · ● Essential A standardised test set or task suite used for comparison; it may not predict your production outcomes.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Golden dataset Evaluation and assurance · ◇ Practitioner A curated reference set of test cases with reviewed expected answers or criteria.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Groundedness / faithfulness Evaluation and assurance · ◇ Practitioner How well an answer is supported by or consistent with supplied evidence; definitions vary by evaluator.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Precision / recall Evaluation and assurance · ◇ Practitioner Precision measures the relevant share of selected items; recall measures the retrieved share of all relevant items.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Task success Evaluation and assurance · ● Essential Whether the complete user goal was achieved within constraints.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Tool accuracy Evaluation and assurance · ◇ Practitioner Whether the agent selected the appropriate tool and supplied correct arguments.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. LLM-as-judge Evaluation and assurance · ◇ Practitioner Using a model to score outputs; calibrate against human review and check judge bias.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Human evaluation Evaluation and assurance · ● Essential People review outputs using defined criteria; agreement and reviewer expertise matter.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Regression testing Evaluation and assurance · ◇ Practitioner Checking that changes do not break previously acceptable behaviour.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Adversarial evaluation Evaluation and assurance · ◇ Practitioner Testing deliberately difficult or hostile inputs, permissions and tool interactions.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Calibration Evaluation and assurance · △ Advanced Whether expressed or estimated confidence corresponds to observed correctness.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. A/B testing Evaluation and assurance · ◇ Practitioner Comparing alternative systems or configurations on controlled groups or traffic.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Data / concept drift Evaluation and assurance · ◇ Practitioner Changes in input distributions or in relationships between inputs and outcomes.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. Model risk management Evaluation and assurance · ● Essential Identifying, validating, monitoring and governing risks from model use.Enterprise context: A team tests answers, permissions and actions against representative workplace scenarios. MLOps Production and performance · ◇ Practitioner Machine learning operations Practices for building, deploying, monitoring and maintaining ML systems.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. LLMOps Production and performance · ◇ Practitioner Large language model operations Operational practices for LLM applications, including prompts, retrieval, evaluation and deployment.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. AgentOps Production and performance · ◇ Practitioner Agent operations An emerging umbrella term for operating agents: traces, tools, state, quality, cost and controls.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Observability Production and performance · ● Essential Understanding system behaviour through logs, metrics, traces and other signals.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Tracing Production and performance · ◇ Practitioner Recording the sequence of model calls, retrieval, tools and other steps in a task.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Latency Production and performance · ● Essential Time to respond or complete a task; measure user-visible end-to-end time.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. TTFT Production and performance · ◇ Practitioner Time to first token Delay before the first generated token reaches the user.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Throughput Production and performance · ◇ Practitioner Work completed per unit time, such as tokens or requests per second.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Rate limits Production and performance · ● Essential Restrictions on request or token volume over a period.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Streaming Production and performance · ◇ Practitioner Delivering output incrementally as it is generated.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Batch inference Production and performance · ◇ Practitioner Processing groups of requests, often with different cost and turnaround characteristics.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Model routing Production and performance · ◇ Practitioner Selecting models based on task, risk, capability, cost or latency requirements.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Fallback Production and performance · ◇ Practitioner An alternative model, workflow or human route when the primary path fails.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Semantic cache Production and performance · △ Advanced Reusing answers for meaningfully similar queries; requires freshness and permission safeguards.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. GPU / accelerator Production and performance · ◇ Practitioner Hardware used to speed model training or inference.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Edge / on-device AI Production and performance · ◇ Practitioner Running AI close to the user or device, potentially reducing network reliance and data transfer.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Versioning / rollback Production and performance · ● Essential Tracking changes and restoring an earlier model, prompt, index or configuration when necessary.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. Kill switch Production and performance · ● Essential A control that stops an agent or disables actions when intervention is needed.Enterprise context: An operations team investigates slow responses, failed tool calls and model-version regressions. AI FinOps Economics and enterprise adoption · ● Essential Financial accountability and optimisation for AI consumption linked to business value.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Token pricing Economics and enterprise adoption · ● Essential Charges for processed or generated tokens; input, output, cached and reasoning usage may differ by service.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Cost per successful task Economics and enterprise adoption · ● Essential Total relevant cost divided by successful outcomes, including retries and human review.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. TCO Economics and enterprise adoption · ● Essential Total cost of ownership Full lifecycle costs including licences, integration, infrastructure, operations, governance and people.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. ROI Economics and enterprise adoption · ● Essential Return on investment Benefits compared with investment, using explicit assumptions and an appropriate time horizon.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Consumption vs licence Economics and enterprise adoption · ● Essential Usage-based charges / subscription or entitlement charges; deployments may combine both.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Budget / quota Economics and enterprise adoption · ● Essential A spending allowance / a usage or capacity allocation; an alert is not necessarily an enforced cap.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. AI CoE Economics and enterprise adoption · ● Essential AI centre of excellence A coordinating function for standards, expertise, delivery and governance, with defined decision rights.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Use-case portfolio Economics and enterprise adoption · ● Essential A managed set of AI opportunities prioritised by value, feasibility and risk.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. AI readiness Economics and enterprise adoption · ● Essential Preparedness across data, technology, security, skills, governance and operating processes.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. AI literacy Economics and enterprise adoption · ● Essential The ability to understand, use and critically assess AI in relevant work.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Adoption vs value Economics and enterprise adoption · ● Essential Usage measures engagement; outcome measures establish whether AI improves the work.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Change management Economics and enterprise adoption · ● Essential Preparing people and processes for new ways of working and sustained adoption.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Human oversight Economics and enterprise adoption · ● Essential Designated monitoring, intervention and accountability; may extend beyond individual approval steps.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Build / buy Economics and enterprise adoption · ● Essential Choosing custom development, purchased capability or a combination based on lifecycle needs.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Vendor lock-in Economics and enterprise adoption · ● Essential Dependency that makes changing providers difficult due to interfaces, data, skills or commercial terms.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. Pilot vs production Economics and enterprise adoption · ● Essential A limited trial / an operational service requiring ownership, support, security and assurance.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience. AI service ownership Economics and enterprise adoption · ● Essential Accountability for service outcomes, lifecycle, controls, operations, costs and continuous improvement.Enterprise context: A service owner measures cost per successfully resolved request alongside employee experience.
No concepts match. Try another term or category.
Concepts that are often confused RAG / fine-tuning Retrieve knowledge at runtime / change model weights through training. MCP / A2A Connect AI applications to tools and context / enable communication between agents. Workflow / agent Predefined paths / model-directed next steps within boundaries. Context / memory Information available now / information retained for later use. Responsible AI / guardrails Lifecycle principles and accountability / specific controls. Open weights / open source Available model parameters / broader openness and licensing requirements. Grounded / correct Supported by sources / factually accurate. Sources can be wrong. Emerging concepts and sources Context engineering, agent harnesses, agentic RAG, MCP, A2A, computer use, test-time compute and AgentOps are important evolving areas. This is a dated reference, not a live news feed. Product and protocol versions require separate checks. AgentOps and agentic AI have varying industry usage; AGI and ASI definitions remain contested.