AI & SaaS glossary

An AI program able to chain actions autonomously to complete a task: search, use tools, decide.

A standardised entry point that lets two software systems exchange data and communicate with each other.

A set of techniques that let a machine simulate human abilities such as understanding, reasoning or deciding from data.

A program that converses in natural language with users, often powered by an LLM, to answer questions or trigger actions.

On-demand delivery of computing resources (servers, storage, software) over the Internet.

Technologies that let a machine analyse and interpret images or video: document reading, quality control…

An AI assistant embedded in work tools (code, office) that suggests, completes or drafts from context.

A numerical representation of text (or images) that captures meaning, used to compare and retrieve content by similarity.

Retraining an existing model on specific data to adapt it to a business domain or a given tone.

An AI model's answer that looks plausible but is actually wrong or made up.

An AI model trained on vast text corpora to understand and generate natural language: answers, summaries, writing.

A branch of AI where a model learns rules from examples, rather than being explicitly programmed.

Monthly (MRR) or annual (ARR) subscription revenue: key indicators of a SaaS model's health.

An architecture where a single software instance serves several customers while strictly isolating their data.

The field of AI dedicated to analysing and generating human language, in both text and voice.

Technology that converts a scanned document or image into usable, searchable text.

The journey that guides a new user or customer until they are self-sufficient with the software.

A natural-language instruction sent to an AI model to obtain a specific result.

A technique that feeds a language model with your own documents to produce reliable, sourced answers.

Software hosted in the cloud and accessed by subscription through a browser, with no client-side install or maintenance.

A contractual commitment on service quality, notably uptime and support response times.

A unit of text (a word fragment) a language model reads or generates; also the billing unit of AI APIs.

A database optimised to store embeddings and retrieve the closest content — the technical backbone of RAG.

An automatic, real-time notification sent from one application to another when an event occurs.