← Home

SkillUp AI · Session 1 of 6

AI for Everyone

What You Actually Need to Know

90 minutes · Interactive · Zero prerequisites

Today's Journey

1The AI Hierarchy — Demystifying the buzzwords15 min
2How GenAI Works — Tokens, prediction, temperature20 min
3What AI Can Do — Real-world use cases15 min
4AI on AWS — Services that matter10 min
5Prompt Tips — A taste of prompting10 min
6PartyRock Exercise — Build your first AI app15 min
7Wrap-Up & Next Steps5 min

Before We Start…

🗳️ Quick Poll

"Which of these have you used?"

ChatGPT  ·  Claude  ·  Gemini  ·  Kiro  ·  None

Confidence Check

📊 Rate yourself 1–5

"How confident are you explaining AI to a colleague?"

1 = No idea  →  5 = Could teach it

🎯 Goal: By the end, you'll be able to explain AI to anyone.

Part 1

The AI Hierarchy

Demystifying the Buzzwords

Four Layers of AI

🔵  Artificial Intelligence — Machines that mimic human intelligence

🟣  Machine Learning — AI that learns from data (no hard-coded rules)

🟡  Deep Learning — ML using neural networks (inspired by the brain)

 Generative AI — DL that creates new content (text, images, code)

GenAI is what's new — it's the reason for today's AI wave

ASEAN Examples

LayerExample
AISpam filters, voice assistants (Siri, Alexa)
MLGrab's route optimization, fraud detection at banks
DLFace recognition on your phone, medical image analysis
GenAIChatGPT writing emails, Claude answering questions, Midjourney images

🍳 The Cooking Analogy

AI = All of cooking

ML = A specific technique (like stir-frying)

DL = A sophisticated version (mastering wok hei)

GenAI = A chef that invents entirely new recipes

Part 2

How GenAI Actually Works

The 5-Minute Version

4 Steps to Generate Text

📖

1. Training

AI reads billions of pages — learns patterns in language

🧩

2. Tokenization

Text splits into "tokens" — chunks, not whole words

🎯

3. Prediction

Predicts the most likely next token — one at a time

4. Generation

Repeat step 3 thousands of times → paragraphs appear

🧩 Tokens — How AI Reads

AI doesn't read words — it reads tokens

"Understanding"Under + stand + ing = 3 tokens

"AI"AI = 1 token

"สวัสดีครับ"วัสดีครับ = ~4 tokens

💰 Cost insight: AI pricing is per-token. Thai uses ~3x more tokens than English for the same meaning — multilingual content costs more!

🎯 Next-Token Prediction

The world's most sophisticated autocomplete

"The cat sat on the ___"

mat (45%) · floor (22%) · chair (15%) · table (8%) · ...

AI picks the most likely next token, adds it, then predicts again.
This is how it writes paragraphs, code, and poems.

🤔 Why Hallucinations Happen

⚠️ AI doesn't "understand" like humans. It predicts what sounds right based on patterns — not what IS right.

→ That's why AI sounds confident even when wrong

→ It's not lying — it's predicting tokens that pattern-match well

Always verify: numbers, dates, names, specific claims

🌡️ Temperature — The Creativity Dial

🧊

Low (0.0–0.3)

Factual, predictable

Use for: compliance reports, data extraction, summaries

⚖️

Medium (0.5–0.7)

Balanced, reliable

Use for: emails, reports, general Q&A

🔥

High (0.8–1.0)

Creative, unreliable

Use for: brainstorming only — verify everything

👔 Analogy: Low = your cautious compliance officer · Medium = reliable PM · High = creative strategist

Part 3

What AI Can Do Today

Real-World Use Cases for Every Role

AI Can…

CapabilityExampleFor
📋 Summarize30-page report → key bullet pointsAll roles
✍️ DraftFirst-draft emails, proposals, SOPsAll roles
🗂️ ClassifySort: invoice / receipt / contractOps, Finance
🔍 ExtractPull names, amounts, dates from docsFinance
🌏 TranslateContent across ASEAN languagesRegional
💻 CodeGenerate scripts, automate tasksTech
📊 AnalyzeIdentify patterns, anomalies, trendsAnalytics

⚠️ What AI Can't Do (Yet)

  • Guarantee facts — Always verify numbers and dates
  • Replace judgment — Complex decisions need human context
  • Access your systems — Needs explicit integration
  • Learn from you — Each session starts fresh

🎯 Rule of thumb: Use AI for first drafts and pattern recognition. Use humans for final decisions and verification.

Part 4

AI on AWS

The Services That Matter

4 Services to Know

🧠

Amazon Bedrock

Access to top AI models (Claude, Titan, Llama) via API

💬

Amazon Q Business

AI assistant that answers from your company's data

👨‍💻

Kiro

AI-powered development environment — specs, steering, agentic workflows

🎉

PartyRock

No-code AI app builder — free, no AWS account needed

You don't need to build these — just know they exist so you can ask the right questions.

Part 5

3 Tips to Talk to AI Better

A Taste of Prompt Engineering

The Problem

Most people talk to AI like a search engine

→ And get bad results

🔑 The fix: Talk to AI like you're briefing a smart colleague who's never worked at your company.

Tip 1: Be Specific

❌ "Write about Q3 results"

✅ "Write a 3-bullet executive summary of Q3 results for our board deck. Keep it under 100 words. Focus on revenue growth and market expansion in SEA."

Tip 2: Give Context

❌ "Summarize this document"

✅ "I'm a finance manager preparing for a leadership review. Summarize this audit report into 5 key findings. Flag anything needing immediate action."

Tip 3: Iterate, Don't Restart

"Make it shorter" →

"Add bullet points" →

"Make the tone more formal" →

"Add a call-to-action at the end"

AI gets better in conversation. Don't start over — refine.

Part 6 — Hands-On

🎉 PartyRock Exercise

Build Your First AI App — No Code, No Cost

What is PartyRock?

A free playground from AWS to explore AI by building apps.

  • No coding required
  • No AWS account needed
  • No cost — free credits included
  • Build, share, and remix AI-powered apps

partyrock.aws

How It Works

1️⃣  Tell Whiskers what you need — describe your app in plain language

2️⃣  Watch it build — input fields, AI outputs, everything wired together

3️⃣  Try it & iterate — adjust through conversation with Whiskers

4️⃣  Share it — anyone with the link can use your app instantly

🏗️ Your Challenge

Build an AI-powered app for your job

Go to partyrock.aws → Click "Create an App"

Ideas to try:

  • Meeting summary generator
  • Email tone adjuster (casual ↔ formal)
  • Weekly report builder
  • ASEAN language translator with context
  • Interview question generator for your role

💡 Pro Tips for PartyRock

Be specific — "Create a tool that generates 3 email reply options: formal, friendly, and brief"

Add constraints — "Keep outputs under 100 words" or "Use bullet points"

Give context — "This is for a finance team in Southeast Asia"

Iterate with Whiskers — "Add a dropdown for tone" or "Make the output include a subject line"

💬 Chat Challenge

🎤 Share in the chat

"What app did you build (or want to build)?"

We'll react to the most creative ones live!

Wrap-Up

What You Learned Today

Key Takeaways

AI → ML → DL → GenAI — you now know the hierarchy

Next-token prediction — AI is fluent but not always factual

Temperature — low for facts, high for brainstorming

AI can summarize, draft, classify, extract, translate, code, analyze

PartyRock — you can build AI apps with zero code

Prompt tips — be specific, give context, iterate

📝 Your Action Item

This week, try ONE AI task at work:

Summarize meeting notes · Draft an email · Ask AI to explain something · Build a PartyRock app

🎓 Continue on Skill Builder

CourseDurationType
AI Practitioner Learning Plan8 hoursLearning Plan
AI Practitioner Power Hour6 hoursLearning Plan
Fundamentals of Generative AI1.5 hoursDigital Course
Machine Learning Basics30 minDigital Course
GenAI Fundamentals — Terminology10 minQuick Start
Cloud Quest: GenAI Practitioner10 hoursGamified

All free on skillbuilder.aws · Requires active Skill Builder account

Coming Next…

Session 2: Talk to AI

Live Prompt Engineering Workshop

Hands-on techniques · Chain-of-Thought · Personas · Structured Outputs · AI Judge Competition

Thank You! 🙏

Interactive explainer: session1-ai-for-everyone.html

PartyRock: partyrock.aws · Skill Builder: skillbuilder.aws