Dictionary
Definition layer. The vocabulary of the system.
AEO (Answer Engine Optimization)
Optimizing content to be cited by AI answer engines like ChatGPT, Perplexity and Google AI Overviews.
Affiliate Marketing
Earning commissions by recommending other companies' products through trackable links.
Agent Architecture
Agent architecture is the structural blueprint of an AI agent: the model, the planning loop, the tools, the memory layer and the guardrails that decide how it acts.
Agent Cost Control
Agent cost control is the practice of budgeting tokens, steps and model tiers per task so autonomous systems stay economically viable at scale.
Agent Handoff
Agent handoff is the controlled transfer of a task — with its context — from one agent to another agent or to a human.
Agent Memory
Persistent context that lets agents retain preferences, decisions, and prior work.
Agent Planning (ReAct, Plan-and-Execute)
Agent planning is how an AI agent decides its next step — reactively step-by-step (ReAct) or by drafting a full plan up front (plan-and-execute).
Agentic RAG
RAG where an agent decides what to retrieve, when, and from which source — instead of a single static query.
Agentic Workflow
A workflow where an LLM decides the next step instead of a hard-coded path.
AI Agent
An autonomous AI system that plans and executes multi-step tasks.
AI Content Pipeline
An end-to-end system that takes a topic and outputs publish-ready content.
AI Copilot
An in-product AI assistant that helps a user complete a task inside an existing workflow.
AI Cost Control
AI cost control is the practice of monitoring, analyzing, and managing the financial expenditures associated with developing, deploying, and operating artificial intelligence systems.
AI Employee
An AI employee is a persistent agent that owns a defined role — with a job description, tools, memory, KPIs and a manager — instead of running as a one-off task.
AI Evals
Reproducible test suites that measure LLM output quality across model, prompt and code changes.
AI Evaluation
AI evaluation is the measurement layer of an AI system: a fixed set of cases, a scoring method and a tracked pass rate you can regress against.
AI Governance
AI governance is the set of policies, records and reviews that make an organisation's AI use accountable and auditable.
AI Governance Framework
An AI Governance Framework is a structured system of policies, processes, roles, and standards designed to guide the responsible, ethical, and compliant development and deployment of artificial intelligence systems within an organization.
AI Monitoring
AI monitoring is production observability for model-driven systems: traces, cost, latency, tool failures and output-quality drift.
AI Orchestration
Coordinating multiple AI models, tools and steps into a single reliable workflow.
AI Router
A layer that picks the cheapest capable model for each request, saving cost and latency.
AI SDR (Sales Development Rep)
An autonomous system that researches, qualifies, and contacts leads.
AI Security
AI security protects systems where the model is an untrusted decision-maker acting on untrusted input with real tool access.
AI Software Engineering
AI software engineering is the practice of building software where agents write most of the code and humans own architecture, review and verification.
AI Testing
AI testing covers two things: using AI to generate and maintain tests, and testing AI systems whose output is non-deterministic.
AI Voice Agent
An AI voice agent is a software program that interacts with users using natural language spoken input and output, performing tasks or providing information. These agents leverage technologies like Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) to simulate human-like conversations.
AI Voice Agent Latency
AI voice agent latency refers to the delay between a user speaking and an AI voice agent's response, critically impacting the naturalness and effectiveness of real-time voice interactions.
API
A defined contract that lets one program call another over the network.
AutoGen Framework
AutoGen is a framework from Microsoft that enables the development of multi-agent conversation systems, allowing agents to communicate, negotiate, and collaborate to achieve complex tasks through dynamic dialogue.
Automatic Speech Recognition (ASR)
Automatic Speech Recognition (ASR) is a technology that converts spoken language into written text, acting as a core component for voice assistants, dictation software, and transcription services. It enables machines to understand human speech.
Automation Observability
Monitoring inputs, model calls, outputs, cost, latency, and failures across AI workflows.
Autonomous Workflow
An autonomous workflow runs end-to-end without a human triggering each step — an agent decides the path, while humans set goals and approve exceptions.
BM25
The keyword-ranking algorithm behind classic search engines.
Browser Agent
A browser agent drives a real web browser — navigating, clicking, typing and reading pages — to complete tasks on sites that have no API.
Canonical URL
The single URL a search engine should credit for a piece of content.
Chain of Thought
Prompting an LLM to reason step-by-step before answering, often improving accuracy on hard tasks.
Chain-of-Thought Prompting
Instructing a model to think step by step before answering.
Chunk Size
The length of text segments stored in a vector index.
Chunking
Splitting documents into retrievable pieces before embedding them for RAG.
Claude Code
Anthropic's terminal-native coding agent that edits real codebases.
Coding Agent
A coding agent reads a repository, plans a change, edits files, runs tests and iterates until the task passes — instead of just suggesting snippets.
Computer-Use Agent
An AI agent that controls a desktop or browser via screenshots and clicks.
Context Caching
Reusing pre-processed input across requests to cut cost and latency.
Context Engineering
Context engineering is the discipline of deciding exactly what information enters a model's context window, in what order and at what cost.
Context Window
The maximum amount of text (in tokens) an LLM can consider in a single call.
Contextual Compression
Contextual compression is a technique used to reduce the size of the input context for a Large Language Model (LLM) while retaining its most relevant information, typically by summarizing or filtering.
Cosine Similarity
The dot-product-of-unit-vectors metric that ranks embeddings.
Cost Per Token
The unit economics of LLM APIs.
CrewAI Framework
CrewAI is a framework for orchestrating role-playing autonomous AI agents, enabling them to collaborate, delegate, and execute tasks as a team. It focuses on clearly defined roles, tasks, and a shared goal to build highly collaborative AI systems.
Data Residency Compliance
Data Residency Compliance refers to the legal requirement for data, particularly personal or sensitive data, to be stored and processed within specific geographic boundaries, typically a country or region.
Deep Research (AI)
Long-running AI research task that produces a cited multi-page report.
Document Extraction Agent
A document extraction agent reads unstructured files — PDFs, scans, emails — and returns validated structured data.
DPO (Direct Preference Optimization)
A simpler alternative to RLHF that skips the reward model.
Edge Computing
Running code and AI inference close to the user instead of in a central data center.
Embedding
A numerical vector representation of text, image or audio that captures meaning for similarity search.
Embedding Dimensions
The vector length of an embedding model's output.
Entity Extraction
Pulling structured entities (people, places, orgs, dates) from text.
Entity SEO
Optimizing around real-world entities (people, products, concepts) instead of just keywords.
Episodic Memory
Session-scoped memory of specific past events.
EU AI Act Compliance
EU AI Act Compliance refers to adhering to the regulatory framework established by the European Union to govern the development, deployment, and use of artificial intelligence systems within the EU.
Evals
Automated tests that grade LLM outputs against expected behavior.
Few-Shot Prompting
Showing the model 2–5 examples of the task inside the prompt so it mirrors the pattern.
Fine-Tuning
Continuing to train a base model on your own examples to specialize its behavior.
Function Calling
The API mechanism that lets a model request a tool call.
Generative Engine Optimization (GEO)
Optimizing content to be cited by generative answer engines like ChatGPT, Perplexity and Google AI Overviews.
Guardrails
Runtime checks that constrain LLM inputs and outputs to keep behavior safe and on-spec.
Hallucination
When an LLM confidently generates false or fabricated information.
Human-in-the-Loop
A control pattern where humans review high-risk AI decisions before execution.
Hybrid Search
Combining keyword (BM25) and vector search for better recall.
Idempotency
The property that running the same operation twice produces the same result as running it once.
Inference
Running a trained model to produce outputs.
Internal Linking
The graph of links between pages on the same domain — the strongest lever most sites underuse.
Jailbreak
A prompt crafted to bypass a model's safety alignment.
JSON Mode
A model setting that guarantees valid JSON output.
Knowledge Graph
A network of entities and the relationships between them, queryable like a map.
LangGraph Framework
LangGraph is a Python library built on LangChain that enables building stateful, multi-actor applications with LLMs by modeling agentic workflows as graphs. It allows for defining complex agent behaviors, including loops and conditional logic, crucial for advanced AI agent orchestration.
LLM (Large Language Model)
A model trained on huge text corpora that predicts the next token to produce human-like language.
LLM Observability
Tracing every prompt, tool call, and token in production.
LLM Orchestration
Coordinating multiple model calls, tools, and data sources into one reliable system.
LLM-as-Judge
Using a strong model to grade another model's output.
llms.txt
A markdown index at /llms.txt that helps LLMs read your site.
Long-Term Memory
Persistent state an agent carries between sessions.
LoRA
Low-Rank Adaptation: cheap fine-tuning that patches only small matrices.
MCP (Model Context Protocol)
Open protocol that lets LLMs connect to tools, data sources and apps through a standard interface.
MCP (Model Context Protocol)
Open standard letting AI clients call external tools, data and prompts.
MCP Client
An MCP client is the AI-side host that discovers servers, lists their tools and routes the model's calls to them.
MCP Resources
MCP resources are addressable read-only context — files, records, docs — that a client can pull into the model instead of calling a tool.
MCP Server
An MCP server exposes tools, resources and prompts from one system so any MCP-compatible AI client can use them over a standard protocol.
MCP Tools
MCP tools are typed, described functions an AI model can call — the unit of capability that decides whether an agent is useful or dangerous.
MCP Transport (stdio vs Streamable HTTP)
MCP runs either as a local stdio process or as a remote streamable HTTP endpoint — the choice decides auth, deployment and who can reach it.
Model Distillation
Training a small model to mimic a large one.
Model Routing
Sending each request to the cheapest model that can handle it.
Moderation
Filtering unsafe input or output before it reaches users.
Multi-Agent System
A multi-agent system splits a job across several specialised AI agents that coordinate through a shared plan, message bus or orchestrator.
Multimodal AI
Models that natively process more than one input type — text, images, audio, or video.
Multimodal Model
A model that reads and reasons across text, images, audio, and video.
n8n
Open-source workflow automation you can self-host.
No-Code Automation
Building business workflows visually without writing code.
PII Redaction
Stripping personally identifiable information before sending to a model.
Planner-Executor Pattern
Splitting agents into a planner LLM and a smaller executor LLM.
Programmatic SEO
Generating hundreds or thousands of targeted pages from a structured dataset.
Prompt Chaining
Pipelining LLM calls where each step's output feeds the next.
Prompt Injection
An attack where hostile input hijacks the LLM's instructions, causing it to leak data or misbehave.
Prompt Template
A reusable prompt with named variables filled at runtime.
Prompt Versioning
Treating prompts as code: tracked, diffed, rollback-able.
Provenance
The trace of where a generated artifact came from.
Quantization
Shrinking a model by lowering weight precision.
RAG (Retrieval-Augmented Generation)
Inject external knowledge into an LLM at query time.
Reflection
An agent critiquing and revising its own output before returning.
Reranker
A second-stage model that reorders retrieved chunks for precision.
Reranking
A second-pass model that reorders retrieved chunks by true relevance to the query.
Research Automation
Research automation turns a question into a sourced, structured answer using search, retrieval, extraction and synthesis agents.
Retrieval
Selecting the most relevant chunks for a query before generation.
RLHF
Reinforcement learning from human feedback, the alignment technique behind ChatGPT.
Schema Markup
Structured JSON-LD tags that describe your page to search engines and answer engines.
Self-Consistency
Sampling multiple answers and picking the majority to reduce errors.
Semantic Search
Finding information by meaning rather than exact keyword match.
Sitemap
An XML file listing a site's URLs for search-engine crawlers.
Streaming
Returning tokens as they are generated instead of waiting for completion.
Structured Data
Schema.org JSON-LD that describes page content to machines.
Structured Generation
Constraining decoding to match a schema at every token.
Structured Output
Forcing AI responses into predictable schemas that software can use.
Summarization
Compressing text while preserving meaning and structure.
Synthetic Data
AI-generated data used to train, test or evaluate models when real data is scarce or sensitive.
System Prompt
A high-priority instruction that sets the model's role, tone and constraints for the whole conversation.
Temperature
The randomness knob on an LLM's output distribution.
Text-to-Speech (TTS)
Generating natural-sounding audio from text.
Text-to-Speech (TTS)
Text-to-Speech (TTS) is a technology that converts written text into spoken words, allowing digital devices to vocalize content. It is a fundamental component of AI voice agents, screen readers, and navigation systems.
Token
The atomic unit an LLM reads and generates — roughly ¾ of a word in English.
Token Budget
A token budget is a predefined limit on the number of tokens an AI application or specific request can consume within a given period or for a single interaction. It is a critical mechanism for controlling costs and managing resource allocation for Large Language Model (LLM) usage.
Token Budgeting
Token budgeting is the strategic allocation and management of token usage within large language model (LLM) operations to control costs and optimize performance.
Tool Calling
The model-to-system interface that lets an LLM trigger external actions.
Tool Use
An LLM invoking external functions to act beyond text generation.
Top-p (Nucleus Sampling)
Restricting sampling to the smallest set of tokens whose probability sums to p.
Topical Authority
The signal that a site comprehensively covers a topic, earned by depth + interlinking, not just backlinks.
Transcription (ASR)
Converting speech audio into text.
Vector Database
A database optimized for similarity search over embeddings.
Vector Index
A data structure that makes nearest-neighbor search fast.
Vibe Coding
Building software by describing intent in natural language and letting AI generate the code.
Vision Model
A model that interprets images as first-class input.
Webhook
An HTTP callback that lets one system push data to another the moment an event happens.
Workflow Trigger
The event that starts an automated workflow.
Zero-Shot Prompting
Asking a model to do a task with no examples in the prompt.