Unit 4Advanced Prompt Engineering
0%

SMU Academy · Self-guided unit

From tool to workflow.

You already know how to prompt. This unit turns AI from something you occasionally reach for into part of how you actually work.

The recap you’re building on — give your “smart intern” everything they need: Task · Context · Instructions · Examples

Your journey this unit

Five moves, each one widening what “using AI” means.

    ① Advanced Prompting Techniques

    Six ways to get a far better answer

    Running example throughout the deck: a sales deck for Alccelerate, an AI consultancy pitching skeptical Singapore SMEs.

    Pick a goal to see the technique that fits.
    📋 Cheat-sheet: when to use what
    ② Go Beyond the Chat Box

    ChatGPT is more than a text box

    Running example: planning a market-entry strategy for SME lending in Thailand.

    ③ Go Beyond ChatGPT

    Pick the right model for the job

    Don’t stress about the “best” model — match the model to the task.

    Pick a task type to see a suggested model.

    The CLEAR framework for prompts

    Compare models live: llm-stats playground · Artificial Analysis leaderboard

    ④ Go Beyond AI Chat Bots

    Source-grounded AI & AI inside your apps

    📓 NotebookLM — answers only from your sources

    Upload your materials; it won’t answer beyond them. Great for learning, research, and communication.

    🧩 AI inside Google Workspace

    Use AI right where the work already is.

    ⑤ Go Beyond AI as a Tool

    Turn a repeated prompt into a reusable skill

    A Project or Custom GPT = a master prompt + files + testing. Build one below.

    🛠️ Master Prompt Builder

    Fill the four parts. Your prompt assembles live — copy it into a ChatGPT Project, Custom GPT, or Claude.

    Your master prompt
    
          

    Projects vs Custom GPTs

    🚦 Before you ship it to others

      Products need a far higher bar than internal tools — aim for 100% correct, not 99%, frame it as an experiment, and add a human check.

      Knowledge check

      Six quick questions. Instant feedback.

      Glossary & takeaways

      📖 Glossary

      🎯 The one idea

      Every section is the same move: stop treating AI as a one-off tool and wire it into a workflow — a repeatable skill with context, examples, testing and a human in the loop.