Executive Cheat Sheet
AI Glossary for Business Leaders
The AI terms that actually come up in the boardroom, defined in plain English, the way a leader of a professional-services firm needs them. No jargon, no math.
- Agent
- An AI system that can take actions on your behalf (calling tools, browsing, or completing multi-step tasks) rather than just answering a question.
- Agentic AI
- AI designed to pursue a goal across multiple steps with limited supervision, deciding what to do next and using tools to get there.
- AI Payback Project Strategy for AI
- A deliberately chosen first AI project (high time-savings, low risk) picked so it pays for itself quickly and funds the next phase of adoption.
- AI Whisperer Strategy for AI
- A person who is naturally skilled at getting strong results from AI tools; often your best internal champion to lead adoption.
- API
- Application Programming Interface: a standard way for software systems to talk to each other; how most business tools connect to AI models behind the scenes.
- Context Window
- How much text an AI model can consider at once, its short-term memory. Larger windows let it work with longer documents.
- Fine-Tuning
- Further training a general model on your own examples so it performs better on a specific task or in your firm's voice.
- Generative AI
- AI that creates new content (text, images, code, audio) rather than only classifying or predicting.
- Hallucination
- When an AI states something false or fabricated as if it were true; the main reason human review still matters.
- Inference
- The act of a trained model producing an answer, “running” the model, as opposed to training it.
- Large Language Model (LLM)
- An AI model trained on vast amounts of text to understand and generate human language; the engine behind tools like ChatGPT and Claude.
- Multimodal
- A model that can work with more than one type of input or output, e.g. text, images and audio together.
- Prompt
- The instruction or question you give an AI model to get a result.
- Prompt Engineering
- The practice of writing and refining prompts to get more accurate, useful output.
- RAG (Retrieval-Augmented Generation)
- A technique that lets an AI answer using your own documents (retrieving relevant passages and feeding them to the model), which reduces hallucination.
- Token
- The small chunk of text (roughly three-quarters of a word) that models read and generate; usage and limits are measured in tokens.
- Training
- The process of teaching a model patterns from large datasets so it can perform tasks; expensive and done before the model is used.
- Governance & Guardrails
- The policies, approvals and controls that decide who can use AI for what, and how outputs are checked, so AI is adopted safely.