Chapter 2 Lab: Build Your CRAFT Prompt Library

🔬 Lab Instructions

Chapter 2 Lab moves theory into muscle memory. You will write five CRAFT prompts you'll actually use, ideate AI opportunities across all three HBR categories (per Mr. Giang's review), score each one with the rule-based grid (/10), and write a short report on what you'll do with these skills over the next 3 months.

Lab goal: Walk out with (1) a 5-prompt personal CRAFT library, (2) one HBR-ideation sheet with ≥6 candidate tasks scored, (3) a 90-day skill-utilization report, (4) feedback to Quanskill.

📌 Lab Flow at a Glance

Confirm

3 quizzes — CRAFT, tool tiers, and the data-sensitivity rule.

Build

5-prompt personal library + the ⭐ HBR Ideation Sheet (per Mr. Giang).

Score & Report

Score each idea /10, write 90-day report, give feedback to Quanskill.

📝 Quiz 1: The CRAFT Letters

You write a prompt: "Last Friday Line 3 night shift produced 1,800 syringes with hairline barrel cracks. Cooling fan tripped twice. Act as ISO 13485 RA analyst. Draft a 6-section CAPA report. Vietnamese, 400 words, formal. QA Manager + 2027 auditor will read." Which CRAFT letter is the WEAKEST in this prompt?

A) C — Context

B) R — Role

C) A — Action (it's vague: 'a 6-section CAPA report' doesn't name the sections)

D) F — Format

📝 Quiz 2: Tool Tier Match

MPV's Maintenance Engineer wants to deploy AI-based predictive maintenance on five injection presses, using 6 months of sensor data already collected. Which tool tier is the most natural starting point?

A) Tier 1 — Foundation model (ChatGPT API)

B) Tier 2 — Microsoft Copilot in M365

C) Tier 3 — Vertical AI (FPT AI Factory, Azure ML for manufacturing)

D) Tier 4 — Power Automate

📝 Quiz 3: The Rule of Thumb Before You Paste

An MPV Finance analyst wants to paste this into free ChatGPT to draft variance commentary: "Revenue Apr 2026: 18.2B VND (vs budget 19.5B). Customer X (Bach Mai Hospital): 4.1B. Customer Y (FV Hospital): 2.8B. Top contributors..." What should she do?

A) Paste it — it's just numbers

B) Either use the enterprise AI plan OR desensitise the data (replace customer names, round numbers) before pasting

C) Avoid AI entirely for financial work

Quick Recap — After Quizzes

CRAFT — A=Action must name the verbs Tier 3 vertical AI for sensor + industrial Desensitise OR enterprise before pasting

⭐ STAR EXERCISE — Your HBR Ideation Sheet (22 min)

This is the exercise Mr. Giang's review explicitly asked for. Generate at least 2 candidate AI tasks per HBR category (total ≥6). Score each on the rule-based /10 grid. Write your 80-word reflection. Give feedback to Quanskill.

The 3 HBR Categories (your ideation lens, from Chapter 1):
① Automating Business Processes — repetitive admin/back-office tasks (47% of HBR's 152 projects)
② Gaining Insight Through Data Analysis — pattern detection, prediction, anomaly spotting (38%)
③ Engaging With Customers/Employees — chatbots, NL interfaces (16% — roll out carefully)

Step 1 — Generate ≥2 candidate tasks per HBR category from YOUR real work

① Process Automation candidates (≥2):

② Cognitive Insight candidates (≥2):

③ Cognitive Engagement candidates (≥2):

Step 2 — Score Each Candidate on the Rule-Based Grid (/10)

For each candidate above, give it 0–2 on each of the 5 dimensions, total /10. Be honest — over-optimistic scores are the #1 reason AI pilots fail.

Dimension	0 points	1 point	2 points
Frequency	A few times/year	Weekly	Daily or many/day
Time spent	< 10 min	10–30 min	> 30 min
Data sensitivity	Highly sensitive (needs enterprise AI)	Some sensitive (needs desensitisation)	Mostly non-sensitive (public AI OK)
Output structured	Different every time	Partly templated	Fits clear template
Good-known	Hard to judge correctness	Spot-checkable	Easy to recognise correct answer

🟢 Score 8–10

Strong candidate. Pilot in 2–4 weeks.

🟡 Score 5–7

Revisit — refine scope or wait for Phase 2.

🔴 Score 0–4

Drop — AI unlikely to help here.

Your scoring table (one row per candidate above):

Step 3 — 80-Word Report: What I'll Do With These Skills in 3 Months

This is the "what they'd utilize skills for" piece from Mr. Giang's note. Be specific and concrete.

Step 4 — Feedback to Quanskill

What was useful? What was confusing? What would make Chapter 2 stronger?

📚 Exercise — Your Personal 5-Prompt CRAFT Library (20 min)

The single most-leveraged habit of this course. Write 5 CRAFT prompts you'll genuinely use in week one. Reuse them, refine them, share at least one with a colleague.

CRAFT Prompt #1 — for your highest-frequency task:

CRAFT Prompt #2 — for a second recurring task:

CRAFT Prompt #3 — for a translation or Vietnamese-language task:

CRAFT Prompt #4 — for an analysis / comparison task:

CRAFT Prompt #5 — for a customer or supplier reply you write often:

🛠️ Exercise — Tool Decision for Your Top Opportunity (8 min)

Take your highest-scored opportunity from the Star Exercise. Run it through the four-question framework.

Opportunity I'm taking through the framework:

Q1 — Commodity capability or genuine differentiator?

Q2 — Data sensitivity bucket (4 options from Ch 2.7):

Q3 — How fast do I need it (Days / Weeks / 1–3 months / 3–9 months)?

Q4 — Vendor question for our IT/procurement team:

My tool decision:

Quick Recap — Before the Report

5-prompt CRAFT library built ≥6 ideas scored across 3 HBR categories Top opportunity through 4-Q framework 90-day plan written

📤 Generate Your Chapter 2 Lab Report

Click below to compile your CRAFT library, HBR Ideation Sheet scores, tool decision, and 90-day plan into one document. Save it for your Phase 2 Capstone folder.

✅ Chapter 2 Complete

You have a working CRAFT prompt library, six scored HBR-categorised ideas, a 90-day skill-utilisation plan, and a working tool decision. Chapter 3 makes all of this safe — Responsible AI for MPV.

📖 Chapter 3 Theory →