70825c1a15
Week 1: The AI Multiplier (MCP + Skills) Week 2: Making AI Actually Work (CLAUDE.md) Week 3: Real Stories, Real Results Week 4: What Could You Build? (Interactive) Skeleton decks in presenterm format. PLAN.md has full series architecture, structure, speaker notes, and open decisions.
373 lines
7.4 KiB
Markdown
373 lines
7.4 KiB
Markdown
<!-- column_layout: [1, 8, 1] -->
|
|
<!-- column: 0 -->
|
|
<!-- alignment: left -->
|
|
|
|
<!-- font_size: 5 -->
|
|
<!-- column: 1 -->
|
|
<!-- jump_to_middle -->
|
|
<!-- alignment: center -->
|
|
Making AI Actually Work
|
|
<!-- font_size: 2 -->
|
|
CLAUDE.md and the Art of Instruction
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: Last week: what's possible. This week: how to get it. -->
|
|
<!-- speaker_note: Quick recap: MCP gives AI tools, Skills give AI expertise. Today — how to give AI context about YOU and YOUR work. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
Why AI Disappoints People
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
<!-- column_layout: [1, 1] -->
|
|
<!-- column: 0 -->
|
|
|
|
**What people say:**
|
|
<!-- new_line -->
|
|
"AI doesn't work."
|
|
<!-- new_line -->
|
|
"It hallucinates too much."
|
|
<!-- new_line -->
|
|
"Copilot is useless."
|
|
|
|
<!-- pause -->
|
|
<!-- column: 1 -->
|
|
|
|
**What's actually happening:**
|
|
<!-- new_line -->
|
|
"I gave it no context."
|
|
<!-- new_line -->
|
|
"I asked it to guess."
|
|
<!-- new_line -->
|
|
"I gave it the wrong tool."
|
|
|
|
<!-- reset_layout -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
|
|
Imagine hiring a brilliant consultant and telling them nothing about your company.
|
|
|
|
Then being surprised when their advice is generic.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: This reframe is critical. Bad tool ≠ bad category. But also: bad instructions ≠ bad tool. -->
|
|
<!-- speaker_note: The Copilot → "AI doesn't work" pipeline is real. People blame the category, not the vendor. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
The Three Levels
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
```mermaid +render +width:100%
|
|
graph TD
|
|
S["🏢 System Level<br/>Who is the AI? What org is this?<br/>Guardrails, conventions, culture"]
|
|
P["📁 Project Level<br/>What are we building?<br/>Constraints, decisions, patterns"]
|
|
U["👤 User Level<br/>Who am I?<br/>Preferences, style, context"]
|
|
|
|
S --> P --> U
|
|
|
|
style S fill:#ff6b6b,stroke:#333,color:#fff
|
|
style P fill:#4a9eff,stroke:#333,color:#fff
|
|
style U fill:#51cf66,stroke:#333,color:#fff
|
|
```
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
|
|
Each level compounds. The AI gets better at every layer.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: These aren't abstract. They map directly to files and configuration. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
System Level
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
The organisation-wide context. Who is this AI? What does it know?
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
**"Be helpful"** → generic, useless
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
**"You are an Axway services consultant. You know our deployment patterns use GitOps via ArgoCD. Customer naming follows the CUST-REGION-ENV pattern. Escalation path is engineer → team lead → director. Never modify production without explicit approval."**
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
→ Now the AI has context that would take a new hire weeks to absorb.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: Read the second one aloud. Let it land. That's the difference. -->
|
|
<!-- speaker_note: This is what CLAUDE.md IS — the onboarding document for your AI colleague. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
Project Level
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
**The "new colleague joins the project" test:**
|
|
|
|
What would you put in their first-day doc?
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- incremental_lists: true -->
|
|
|
|
* What are we building and why?
|
|
<!-- new_line -->
|
|
* What's been decided (and what hasn't)?
|
|
<!-- new_line -->
|
|
* What are the constraints?
|
|
<!-- new_line -->
|
|
* Where is everything? (repos, docs, environments)
|
|
<!-- new_line -->
|
|
* What are the gotchas?
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
|
|
Write that document. Put it in CLAUDE.md or AGENTS.md.
|
|
|
|
Your AI reads it every single session.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: The analogy makes it tangible. Everyone knows what onboarding feels like. -->
|
|
<!-- speaker_note: If you wouldn't expect a new hire to succeed without this context, why expect AI to? -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
User Level
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
Personal working style. The compound effect.
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
<!-- column_layout: [1, 1] -->
|
|
<!-- column: 0 -->
|
|
|
|
**Without:**
|
|
<!-- new_line -->
|
|
- Generic tone
|
|
- Over-explains basics
|
|
- Wrong assumptions
|
|
- Feels like talking to a stranger
|
|
|
|
<!-- pause -->
|
|
<!-- column: 1 -->
|
|
|
|
**With:**
|
|
<!-- new_line -->
|
|
- Matches your communication style
|
|
- Knows your role and expertise level
|
|
- Remembers your preferences
|
|
- Feels like working with someone who knows you
|
|
|
|
<!-- reset_layout -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
|
|
System + Project + User = AI that feels like a team member, not a chatbot.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: This is where it gets personal and relatable. -->
|
|
<!-- speaker_note: "Concise answers" vs "walk me through reasoning" — small preference, huge impact over time. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
Intent vs. Instruction
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
<!-- column_layout: [1, 1, 1] -->
|
|
<!-- column: 0 -->
|
|
|
|
**Vague** 😬
|
|
<!-- new_line -->
|
|
"Check if my server
|
|
is secure"
|
|
|
|
<!-- pause -->
|
|
<!-- column: 1 -->
|
|
|
|
**Better** 👍
|
|
<!-- new_line -->
|
|
"Audit SSH config
|
|
against CIS benchmarks,
|
|
prioritise by severity"
|
|
|
|
<!-- pause -->
|
|
<!-- column: 2 -->
|
|
|
|
**Best** 🎯
|
|
<!-- new_line -->
|
|
Write a skill for it.
|
|
AI follows the same
|
|
checklist every time.
|
|
|
|
<!-- reset_layout -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
<!-- font_size: 2 -->
|
|
|
|
Ambiguity is the enemy of automation.
|
|
|
|
If you don't define "safe"… the AI will.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: Callback to Week 1 skills content. The three levels of prompt quality. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 5 -->
|
|
<!-- jump_to_middle -->
|
|
<!-- alignment: center -->
|
|
DEMO
|
|
===
|
|
<!-- font_size: 2 -->
|
|
Same prompt. Two contexts. Watch what changes.
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: Demo 1: Before and after CLAUDE.md — same prompt, dramatic difference -->
|
|
<!-- speaker_note: Demo 2: Build a CLAUDE.md live with the audience for a real project -->
|
|
<!-- speaker_note: PLACEHOLDER: Decide which project/scenario for the live build -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
Your Homework
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- incremental_lists: true -->
|
|
|
|
* **This week:** Write a CLAUDE.md for one project you're working on
|
|
<!-- new_line -->
|
|
* **This week:** Add 3 sentences about your preferences to your AI tool
|
|
<!-- new_line -->
|
|
* **Next time you're frustrated:** Ask "what context was I not giving it?"
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
|
|
These are 10-minute tasks that permanently improve every AI interaction.
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: Keep it concrete and achievable. Three things, this week. -->
|
|
<!-- end_slide -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<!-- font_size: 4 -->
|
|
Next Week
|
|
===
|
|
<!-- font_size: 2 -->
|
|
<!-- new_line -->
|
|
<!-- pause -->
|
|
|
|
Real stories from real people in this room.
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
|
|
What Dhruv built for $2.50. What Gill discovered in her first hour.
|
|
|
|
What's actually changing in how we work.
|
|
|
|
<!-- pause -->
|
|
<!-- new_line -->
|
|
<!-- alignment: center -->
|
|
<!-- font_size: 2 -->
|
|
🎯
|
|
|
|
<!-- no_footer -->
|
|
<!-- speaker_note: Name drop. Create anticipation. People want to hear peer stories. -->
|
|
<!-- end_slide -->
|