← All guides

AI Guide · 15 short modules

Understand how AI tools work.

A plain-English route through AI models, large language models, and the products built around them. Start with the basics, then see how teams make AI features useful and dependable.

Begin the guide →
Part 01

AI foundations

Learn what AI means, how models learn, and where language models fit.

  1. 1Artificial IntelligenceArtificial intelligence is software that performs tasks associated with human abilities. An AI model is the trained component that takes input and produces an output.
  2. 2Machine LearningMachine learning finds patterns in examples and uses them to make predictions or create new content.
  3. 3Model Training vs InferenceTraining uses examples to change a model; inference uses the trained model to answer a new request or make a prediction.
  4. 4Generative AIGenerative AI creates content such as text, images, audio, or code by using patterns learned from examples.
  5. 5Large Language ModelAn LLM is an AI model trained on large amounts of text to produce language, such as answers, summaries, and drafts.
Part 02

Using a model

See how a language model handles a request and why a convincing answer may still be wrong.

  1. 6AI TokensAn AI token is a small piece of text a language model processes. Token counts affect how much text it can handle at once and often determine usage limits and cost.
  2. 7AI PromptA prompt is the request and supporting information an AI model receives. Prompt engineering means shaping that input to get more useful answers.
  3. 8AI HallucinationAn AI hallucination is a convincing answer that contains made-up, false, or unsupported information.
Part 03

Building an AI feature

Understand the software that supplies information, connects tools, and manages actions.

  1. 9Context EngineeringContext engineering is choosing and organising the information an AI model receives for a request, such as instructions, past messages, documents, and available tools.
  2. 10Retrieval-Augmented GenerationRAG looks up relevant information in documents or databases and gives it to an AI model to use in its response.
  3. 11AI Tool CallingTool calling lets an AI model ask its app to use another product or service. The app checks the request, makes the call, and passes back the result.
  4. 12AI AgentAn AI agent is software that uses a model to choose steps toward a goal, then uses tools to carry them out and check the results.
Part 04

Making AI dependable

Explore the checks, controls, and human decisions that make a system useful in practice.

  1. 13AI EvaluationAI evaluation tests an AI product on realistic tasks to see how well it works, what goes wrong, and whether changes improve it.
  2. 14AI GuardrailsAI guardrails are rules and checks that limit what an AI product can see, say, or do, such as restricting private data or requiring approval for an action.
  3. 15Human in the LoopHuman in the loop means a person checks, corrects, or approves an AI system’s output or action at a defined point.