Software and AI glossary
Short definitions you can quote, plus how each idea works and where it shows up in our hubs. Built for searchers who want clarity without hype.
- Agentic AI
Agentic AI refers to AI systems that plan multi-step work and take actions through tools toward a goal, usually with constraints and oversight.
- AI agent
An AI agent is software that pursues a defined job by choosing tools, reading state and producing a scorable outcome - not a FAQ chatbot alone.
- API
An API is a documented contract that lets one system request actions or data from another in a structured way.
- Embeddings
Embeddings are numeric vectors that represent text (or other media) so similar meanings are close in vector space.
- Evaluation set
An evaluation set is a curated pack of real inputs and expected outcomes used to score AI behaviour over time.
- Fine-tuning
Fine-tuning further trains a base model on labelled examples so behaviour fits a narrower domain or format.
- Human-in-the-loop
Human-in-the-loop means a person reviews or approves AI outputs or actions before irreversible effects occur.
- Large language model (LLM)
A large language model is a neural model trained on text to predict and generate language, often used as a reasoning component inside apps and agents.
- Model Context Protocol (MCP)
Model Context Protocol (MCP) is a convention for exposing tools and contextual resources to models through structured servers.
- Multi-agent system
A multi-agent system is a set of agents that collaborate or hand off work, each with a role, tools and boundaries.
- Multi-tenancy
Multi-tenancy means many customers share infrastructure while their data and configuration stay isolated.
- MVP
An MVP is the smallest product that can test a critical assumption while remaining operable for real users.
- Platform engineering
Platform engineering builds shared internal platforms - paved roads for deploy, identity, data and observability - that product teams use.
- Product engineering
Product engineering is the practice of taking a product from problem framing through architecture, build, release and operation.
- Prompt engineering
Prompt engineering is the craft of instructions, examples and constraints that steer model outputs toward a task.
- Retrieval-augmented generation (RAG)
RAG retrieves relevant passages from your corpus and feeds them into a model so answers cite current sources instead of memory alone.
- SaaS
SaaS (software as a service) is software delivered over the network, operated by a provider, usually on a subscription.
- Single sign-on (SSO)
Single sign-on lets a user authenticate once and access multiple applications through a trusted identity provider.
- Technical debt
Technical debt is the deferred cost of design or implementation shortcuts that slow future change or raise risk.
- Vector database
A vector database stores embedding vectors and retrieves nearest neighbours for semantic search and RAG pipelines.