Data Intermediate 22 min

The AI Data Analyst workflow — CSV in, management summary out

The problem

You upload a sales CSV and ask for insights. You get eight paragraphs of plausible commentary: growth is strong in the west, the premium segment is underperforming, consider focusing on repeat customers. It reads well.

The workflow
1Understand structure
2Compute descriptives
3Compare periods
4Find anomalies
5Separate finding from explanation
6Draft summary
7Verify three numbers
Send

What you need

File: Q3 sales export, 8,400 rows, apparel wholesale business. Decision: whether to change the west-region incentive structure.

The tool

An assistant's code/data-analysis mode with the file uploaded

The prompt

Subscribers. The full prompt is part of the SMEMinds AI Playbook. See plans.

What comes back

MANAGEMENT SUMMARY — Q3 FY26 sales 1. HEADLINE — Q3 revenue was ₹4.82 crore, down 11.4% against Q2's ₹5.44 crore.

Verify before you use it

Human review required. Check every line below before this leaves your screen.
  • Recompute the headline number and the two largest contributors yourself in Excel. Three numbers, two minutes. If they match, the rest of the computed table is very likely sound.
  • Confirm nothing in sections 1–3 is a hypothesis. 'Down because of competition' would be a serious error at this point in the document.
  • Check the date range from step 1 covers a complete quarter. A truncated export is the most common cause of a spurious decline.
  • Section 6 should feel uncomfortably honest. If it is short or vague, push back — it is where credibility comes from.