10-09-2026, 02:13 PM
My routine for messy client spreadsheets, saves ~2 hours each:
1. Paste a sample of the messy column + ask for the cleaning RULE, not the cleaned data
2. Turn that rule into a formula / regex and apply it locally
3. Ask the model to list edge cases it can see, then check them
Asking for the rule (not the output) keeps your data private and scales to 100k rows. The model is the consultant, not the worker.
1. Paste a sample of the messy column + ask for the cleaning RULE, not the cleaned data
2. Turn that rule into a formula / regex and apply it locally
3. Ask the model to list edge cases it can see, then check them
Asking for the rule (not the output) keeps your data private and scales to 100k rows. The model is the consultant, not the worker.
👍 Like0❤️ Love0😂 Haha0🔥 Fire0👏 Clap0⚡ 0
