Tool: ChatGPT
Time: 5 minutes
Result: A ranked list of customers most likely to churn, with reasons and recommended next steps.
Most family history research ends up buried in disorganized folders or long, unsearchable PDFs. It’s hard to see the big picture when the details are scattered across dozens of documents.
AI Win #5
Today’s win is a fast customer health check. Paste a small export of customer data into ChatGPT and ask it to rank accounts by churn risk. You’ll get a simple watchlist your sales, success, or support team can act on this week.
How To Do It
Export 20–100 customer rows from your CRM, billing system, product analytics tool, or spreadsheet. Include fields like last login, usage trend, support tickets, NPS, renewal date, plan size, and notes if available.
Paste the data into ChatGPT with the prompt below. If you have sensitive data, remove names and use customer IDs instead.
Review the ranked watchlist, then send the top 5 accounts to the right owner with the recommended next action.
Why This Is Useful
Churn risk is usually a pattern, not a single metric. This approach combines behavior, sentiment, timing, and context into one practical view. It also forces the output into decisions: who to contact, why, and what to do next.
Copy / Paste Prompt
You are helping me identify customer churn risk from a small customer data export.
Goal: Create a ranked churn watchlist I can act on this week.
Instructions:
1. Analyze the customer data I paste below.
2. Do not treat any single field as proof of churn. Look for combinations of warning signs.
3. Rank customers by churn risk: High, Medium, or Low.
4. For each High and Medium risk customer, explain the specific signals that led to the rating.
5. Recommend one practical next action for each customer.
6. If important data is missing, say what is missing, but still make the best assessment possible.
7. Keep the output concise and easy to scan.
Use this scoring guidance:
- High risk: multiple negative signals, such as declining usage, recent complaints, low NPS, unresolved tickets, approaching renewal, downgrade language, payment issues, or lack of engagement.
- Medium risk: one or two concerning signals, or unclear engagement before renewal.
- Low risk: stable or growing usage, positive sentiment, recent engagement, few support issues, or no urgent renewal risk.
Return the answer in this format:
Churn Watchlist
| Rank | Customer | Risk | Why they are at risk | Recommended next action | Owner priority |
Then add:
1. Top 3 patterns you noticed across the customer base
2. 3 retention actions we should take this week
3. Any data fields that would improve the assessment next time
Here is the customer data:
[paste customer table, CSV, or notes here]Try it with 25 customer rows before your next retention meeting.
