Prompt engineering: get reliable answers from chat models

11 steps · about 7 h 30 min · Beginner

  1. Writing Effective AI Prompts for Business (guide, Google Workspace, 15 min)

    A short, jargon-free start that introduces the four parts of a prompt: persona, task, context and format.

  2. Prompt engineering best practices for ChatGPT (article, OpenAI Help Center, 10 min)

    Adds the two habits everything else builds on: be specific, then iterate on what the model gives back.

  3. Prompt-Engineering-Guide/guides/prompts-intro.md (guide, DAIR.AI, 20 min)

    Shows how a prompt is built from elements such as instruction, context, input and output format, with before-and-after examples.

  4. Anthropic Prompt Engineering Interactive Tutorial (interactive, Anthropic (GitHub), 2 h)

    Hands-on exercises let you practise clarity, roles, examples and step-by-step reasoning, and see the effect of each.

  5. ChatGPT Prompt Engineering for Developers (video, DeepLearning.AI (Andrew Ng, Isa Fulford), 1 h 40 min)

    Teaches two core principles and an iterate-and-test loop through short lessons on summarizing, inferring, transforming and expanding text.

  6. Prompt engineering overview (doc, Claude Docs (Anthropic), 15 min)

    Explains when prompting is the right fix and lists techniques ordered from most to least broadly effective.

  7. Prompt design strategies (doc, Google AI for Developers, 25 min)

    A second vendor's view on structure, context, and splitting or chaining prompts, so you see which habits work across models.

  8. Prompt engineering (doc, OpenAI, 25 min)

    Covers separating instructions from data with Markdown and XML tags, and versioning prompts like code.

  9. Prompt Engineering Guide (reference, DAIR.AI, 1 h)

    Gives the named techniques in one place, such as few-shot, chain-of-thought and prompt chaining, plus guidance on reliability and bias.

  10. Prompting best practices (doc, Claude Platform Docs (Anthropic), 40 min)

    A deeper reference on clarity, examples, structure and thinking, for tightening prompts you already use.

  11. Effective context engineering for AI agents (article, Anthropic, 20 min)

    Closes by widening the view from the wording of a prompt to curating everything the model sees.