Automate Weekly Sales Reports Using AI

CSV to formatted summary in 30 seconds

The problem

You spend an hour every Monday pulling numbers from a CSV and formatting them into a readable report.

The outcome

A formatted summary with week-over-week comparisons, top performers, and a Slack-ready digest.

What you need

Claude Code installed
A CSV file with sales data (columns: date, rep, deal_name, amount, status)

Step by step

Create your project folder

Create a folder for this automation and navigate to it:

bash

mkdir ~/weekly-sales-report
cd ~/weekly-sales-report

Add your data

Drop your sales CSV into the folder. It should have columns like: date, rep_name, deal_name, amount, status.

Don't have one? Download sample_sales_data.csv to try it out.

Create the CLAUDE.md file

Create a file called CLAUDE.md in the folder with the instructions from the section below. You can copy-paste it directly using the copy button.

Run it

Open your terminal and run:

bash

cd ~/weekly-sales-report
claude

This opens Claude Code with your CLAUDE.md loaded into context. Now tell it what to do:

Generate the weekly sales report from the CSV in this folder.

Claude reads your CSV, follows the instructions in CLAUDE.md, and generates a formatted markdown report. Takes about 30 seconds.

Review and customize

Open the generated report. Want different metrics? Edit the CLAUDE.md and run it again. The instructions are plain English — change whatever you want.

The CLAUDE.md file

Copy this into a file called CLAUDE.md in your project folder.

CLAUDE.md

# Weekly Sales Report

Read the most recent CSV file in this folder.

Create a markdown report that includes:

## Summary
- Total revenue (sum of amount column)
- Number of deals closed (where status = "closed")
- Average deal size
- Number of deals still in pipeline

## Week-over-Week
If a previous report exists in this folder, compare:
- Revenue change ($ and %)
- Deal count change
- Average deal size change

## Top Performers
- Top 3 reps by total revenue
- Include their deal count and average deal size

## Alerts
- Flag any rep whose revenue dropped more than 20% from the previous period
- Flag any deals over $10,000 that are still in pipeline

## Format
- Use markdown tables
- Format currency with $ and commas
- Bold any alerts
- Keep it concise — this goes in a Slack channel

Save as report_[today's date].md