Back

Advanced

Create custom assistants, design multi-step workflows, and understand ethics and automation.

HK$8,000

Detailed 5-day curriculum

  1. Day 1: Custom GPTs & projects

    What is a Custom GPT?

    • A tailored assistant with instructions, knowledge, and capabilities for one job.
    • Meal planner for your family's diet
    • Study helper for specific subjects
    • Client FAQ bot for your business

    Hands-on practice

    • Create a Custom GPT or project: define name, instructions, 3 example questions, and boundaries.

    Knowledge check

    • Custom GPT instructions should include…

    Safety habit

    • Don't upload confidential client files to custom GPTs on consumer plans.

  2. Day 2: Multi-step workflows

    Chain prompts

    • Step 1: Research / brainstorm
    • Step 2: Structure / outline
    • Step 3: Draft / create
    • Step 4: Review / polish

    Hands-on practice

    • Design a 4-step content workflow (e.g. blog post, social campaign, client proposal).

    Analogy or example

    • Tap for workflow template
    • Input → Process → Output → QA. Document each step's prompt in your playbook.

    Knowledge check

    • Why use multi-step instead of one mega-prompt?

    Safety habit

    • Review each step's output before feeding it to the next — errors compound.

  3. Day 3: AI + automation intro

    No-code automation

    • Zapier — connect apps with triggers
    • Make (Integromat) — visual workflows
    • n8n — self-hostable option
    • Pattern: When X happens → Ask AI → Do Y

    Hands-on practice

    • Sketch one automation: trigger (new email), AI step (summarise), action (save to doc).

    Example automations

    • New form submission → AI draft reply
    • Weekly calendar → AI prep brief
    • Invoice received → extract key fields

    Knowledge check

    • Before automating client emails, you should…

    Safety habit

    • Automations that send messages or move money need human approval gates.

  4. Day 4: Data & ethics

    Ethics in practice

    • Bias — AI reflects training data biases
    • Copyright — who owns AI output?
    • HK context — PDPO, consumer protection
    • Client data — confidentiality agreements

    Hands-on practice

    • Discuss 2 case studies: What went wrong? What would you do differently?

    Responsible use checklist

    • Disclose AI use when required
    • Don't impersonate real people
    • Respect intellectual property
    • Keep audit trail for business work

    Knowledge check

    • Using AI for client work in HK, you should…

    Safety habit

    • When serving clients, treat their data as confidential — use enterprise tools with data protection when needed.

  5. Day 5: Building your AI playbook

    Your AI operating manual

    • Tools you use and why
    • Standard prompts for repeat tasks
    • Safety rules for you and your household/business
    • When to escalate to a human expert

    Hands-on practice

    • Draft your personal AI playbook — at least 5 SOPs for tasks you do regularly.

    Take-home

    • Finish one working custom assistant and document its workflow in the portal.

    Knowledge check

    • An AI playbook helps you…

    Safety habit

    • Review and update your playbook every 3 months — AI tools change quickly.
Book this moduleSee dates