You're going to work with a real CSV. Before you ask what AI insight is, you have to check if the data and the calculations are reliable enough to generate insight.
20 to 25 minutes.I have a real CSV to download.🧠 Check → Calculate → Interpret → Report🎯 Artifact: Management Performance Review
0% completed
📥 SOURCE FILE
1. Download the actual data file
The file contains performance of the HCM stores. It was intentionally designed to resemble a real job export: not completely clean.
📄
viie-edu-lesson-8-retail-performance.csv
Sales, margin, traffic · LY vs TY · have some data-quality issues that are intentionally installed.
Your workflow today
1 · CHECK Are the data ready to be computed?
2 · CALCULATE Are the formula and denominator correct?
3 · INTERPRET What is a result, what is an explanation?
🎭 MANAGEMENT QUESTION
Which store is best?
Let's say AI sees a store with the highest sales growth and concludes: "This is the best-performing store".
How should you react?
🔎 STEP 1 · DATA QUALITY CHECK
Before you analyze, look for things that could make the wrong calculations.
Open CSV with Excel or upload to ChatGPT.
What's wrong with the file?
🛠 HANDS-ON · PROMPT 1
4. Detract data checks from analysis
Check this CSV file before analyzing performance. Check: - duplicate rows / duplicate keys; - missing values; - unit inconsistency; - denominator equal to 0; - unusual value that may cause incorrect calculations.
Here's Lesson 6 that's applied to the data: AI check → human confirm → is new to AI calculate.
🧮 STEP 2 · CALCULATION CHECK
Fifth, AI is not the business logic.
Margin rate LY
20%
Margin rate TY
18%
What's the best way to communicate in management?
🧮 STEP 2 · DENOMINATOR
6. What is LY = 0 and growth %?
In the file there is a new store: Sales LY = 0, Sales TY = 85 million VND.
🧠 STEP 3 · RESULT VS EXPLANATION
Seven, you're not allowed to jump straight into the sky.
Result
Store A: increased sales, increased traffic, decreased margin.
Explanation
The margin decreases due to deeper promotion.
Which version is more appropriate?
⏸ BUILD CHECKPOINT
Before you use AI to write a report, you need to confirm three things:
1 · Data ready? Duplicate, missing, unit, zero denominator has been dealt with?
2 · Calculation right? The formula, the metric definition and the unit right?
3 · Interpretation calibrated? Fact separated from cause/assumption?
Write down what you won't let AI do with this file.
🛠 STEP 4 · BUILD MANAGEMENT REPORT
After the data is confirmed, report is requested.
Based on the data that has been confirmed, create a Management Performance Review. For each store when the calculation is valid: - Sales growth vs LY - Margin growth vs LY - Traffic growth vs LY Do not count growth % if denominator LY = 0; record new / no comparable LY. Do not self-fill missing values. Do not use sales growth alone to call the store best or worst. Choose up to 3 notable stores. For each store, write: 1. What happened 2. Why Evidence 3. It matters 4. What is still uncertain 5.
🔍 STEP 5 · REVIEW REPORT
10. What better report to the boss?
The art of the body
11. Your Management Performance Review
After running the CSV through the workflow, fill in the most important parts of the artifact you're going to use.
Data Quality Note
KPI / Calculation Note
Stores to Watch
Executive Summary
0/6
You have completed Build 2 · Management Report.
One sentence to carry: Don't start with insight. Start with making sure that data and calculation are reliable enough to generate insight.