AI glossary

AI agent terms, explained plainly.

Short, accurate definitions of the terms you'll meet when planning AI agents and automation, from people who build them every week.

Agentic AI

Agentic AI describes AI systems that plan and carry out multi-step tasks with some autonomy, choosing actions and tools to reach a goal, rather than only generating a single answer to a prompt.

Read definition

An AI agent is software that uses a large language model to pursue a goal on its own: it understands a request, decides which steps to take, uses tools such as calendars, CRMs or databases, and completes the task, asking a human when it is unsure.

Read definition

AI guardrails

AI guardrails are the rules and checks that keep an AI system safe and on-task: limits on topics and actions, validation of outputs, protection of personal data, and automatic escalation to a person in defined situations.

Read definition

AI hallucination

An AI hallucination is an answer from a language model that sounds confident but is false or not supported by its sources. Hallucinations are reduced by grounding answers in trusted data, restricting what the model may claim and letting it say "I don't know".

Read definition

AI voice agent

An AI voice agent is an AI agent that talks on the phone. It turns speech into text, reasons with a large language model, replies in a natural synthetic voice, and can take actions such as booking appointments or updating a CRM during the call.

Read definition

Chatbot

A chatbot is software that holds a text conversation with users. Traditional chatbots follow fixed scripts and buttons; AI chatbots use large language models to understand free-form questions and answer from a knowledge base.

Read definition

Embeddings

Embeddings are lists of numbers that represent the meaning of text, images or other data, so that similar meanings sit close together. They let AI systems search by meaning rather than exact keywords.

Read definition

Human-in-the-loop

Human-in-the-loop (HITL) is a design where people review, approve or correct an AI system's work at defined points, for example checking low-confidence document extractions or taking over a conversation the agent can't resolve.

Read definition

Intelligent document processing (IDP)

Intelligent document processing (IDP) uses OCR and AI models to read documents such as invoices, forms, contracts and IDs, extract the required fields, validate them and pass clean data to business systems, with a human reviewing uncertain cases.

Read definition

IVR (interactive voice response)

IVR (interactive voice response) is a phone menu system that routes callers using keypad presses or simple voice commands ("press 1 for sales"). Unlike an AI voice agent, an IVR follows fixed menus and cannot hold a natural conversation or complete tasks on its own.

Read definition

Large language model (LLM)

A large language model (LLM) is an AI model trained on very large amounts of text to understand and generate language. LLMs such as GPT, Claude and Gemini power chatbots, AI agents, summarisation, extraction and writing tools.

Read definition

Multi-agent system

A multi-agent system is a setup where several specialised AI agents work together, for example one researching, one writing and one checking, coordinated by an orchestrator agent or a workflow, to complete tasks that are too broad for a single agent.

Read definition

n8n

n8n is an open-source workflow automation platform that connects apps and APIs with visual workflows. It can be self-hosted for data control and lower costs, and it includes AI nodes for building agents and LLM steps into automations.

Read definition

OCR (optical character recognition)

OCR (optical character recognition) converts text in images and scanned documents into machine-readable text. It is the first step of most document automation, followed by AI that understands which text belongs to which field.

Read definition

Prompt engineering

Prompt engineering is the practice of writing and structuring instructions, examples and context for a language model so it produces accurate, consistent and safe output for a specific task.

Read definition

Retrieval-augmented generation (RAG)

Retrieval-augmented generation (RAG) is a technique where an AI system first searches a trusted knowledge source, such as company documents or a database, and then gives the relevant passages to a language model so its answer is grounded in that information and can cite sources.

Read definition

Speech-to-text (ASR)

Speech-to-text, also called automatic speech recognition (ASR), converts spoken audio into written text in real time. It is the 'ears' of an AI voice agent and affects how well the agent understands accents, languages and noisy lines.

Read definition

Text-to-speech (TTS)

Text-to-speech (TTS) turns written text into spoken audio. Modern neural TTS voices sound natural, and they are the 'voice' of an AI voice agent, chosen for clarity, tone, language and speed.

Read definition

Tool calling (function calling)

Tool calling, or function calling, lets a language model request an action from software, such as checking a calendar, looking up an order or creating a CRM record, by returning a structured call that the application executes and feeds back to the model.

Read definition

WhatsApp Business API

The WhatsApp Business API (WhatsApp Business Platform) is Meta's official interface that lets businesses connect WhatsApp to software, so AI agents, CRMs and support tools can send and receive messages at scale, with templates for business-initiated messages.

Read definition

Workflow automation

Workflow automation is the use of software to run a sequence of business steps automatically, such as moving data between apps, sending reminders, creating tasks and generating reports, triggered by events like a form submission or a new booking.

Read definition

Free AI audit

Know the terms? Let's build the agent.

A 30-minute call with our engineers. We map your processes, pick the first agent worth building and tell you honestly if AI isn't the answer.

Book a free AI audit

No slides. No obligation. Replies within one business day.