RESOURCES / GLOSSARY
Enterprise AI terms,
defined in plain English.
Short definitions of the terms that come up when enterprises build, buy and check AI systems. Each entry links to a fuller explanation with examples, sources and the questions to ask.
BUILDING AND RUNNING AGENTS
How AI agents work.
Execution agent
An AI agent that carries out a defined task within set limits and returns its result with the sources and calculations behind it.
Read the guide 02Agentic workflow
A business process in which AI agents carry out some steps themselves, using tools and checking results within limits people set.
Read the definition 03Agent orchestration
Coordinating the steps, agents and tools in an AI system, including what happens when a step fails.
Read the definition 04Model Context Protocol (MCP)
An open standard for connecting AI applications to tools and data.
Read the definition 05Grounding
Connecting an AI model’s answer to specific sources, so its claims can be checked against them.
Read the definition 06AI token cost
What a model charges for the text it reads and writes, priced per million tokens.
Use the calculatorTRUST AND CONTROL
How to check and control AI.
Human in the loop
A person reviews, approves or corrects an AI system’s work at defined points.
Read the definition 02AI guardrails
Controls that limit what an AI system can receive, produce and do.
Read the definition 03Agent observability
Tracing what an agent does in production: each step, tool call, cost and error.
Read the definition 04LLM evaluation
Testing a language model or AI agent against defined cases to measure accuracy, safety and cost.
Read the definition 05Data lineage
The record of where data came from, how it moved and how it changed on the way to a result.
Read the definition 06AI provenance
The record of how an AI result was produced: its inputs, the system that made it, who approved it and when.
Read the definition 07AI audit trail
The record of what an agent did: its request, data, steps and result.
See the audit trail 08Auditable AI
An AI system that keeps enough records for a reviewer to check a result after the fact.
Read the guide 09Shadow AI
AI tools or AI-built software used at work without the IT or security team knowing.
Read the guideBUYING AND DELIVERY
How enterprises buy AI.
Services as software
Buying work done mostly by software and AI agents, instead of a tool your team runs.
Read the guide 02Forward deployed engineering
Engineers who work inside a customer’s team to build and run software on the vendor’s platform.
Read the guide 03Agent washing
Marketing a product as an AI agent when it cannot carry out tasks on its own.
Read the definition 04SaaSpocalypse
The 2026 market view that AI agents will replace much of today’s per-seat software.
Read the guide 05Build, buy or extend
The three routes to an enterprise AI application, and the costs each one leaves out.
Read the decision guideHow we write these definitions.
Each entry starts with a one-sentence answer, then explains how the idea works in practice and where it goes wrong. We cite the standard, regulation or research behind each claim, and each page shows when we last revised it.