Count Before You Automate: Tallying Repeated Request Types From an Inbox Export in Python

Описание и применение

Скрипт на Python позволяет быстро подсчитать типы запросов в почтовом ящике, создавая CSV‑отчёт с категориями и примерами, что облегчает аудит автоматизации.

Задача

Owners need accurate counts of recurring request types before automating, but guesses are unreliable.

Результат

The script outputs a CSV with categories, counts, and example subjects, enabling a data‑driven audit of request types.

Рабочий процесс

• Export mailbox or label as .mbox or CSV
• Run Python script that categorises subjects using regex rules
• Collect counts and sample subjects per category
• Iterate on rules to reduce 'other' bucket
• Prepare tally CSV, examples, and notes for audit

Доказательства

Every automation conversation eventually hits the question "how often does this actually happen?" Owners guess. Guesses are usually wrong in both directions.

Источник

If you are preparing for an AI workflow audit with real request counts you can answer that question in an afternoon with an export of your inbox and a short script. Everything runs locally.

Источник

The script below handles mbox; adapting it for CSV is a few lines.

Источник

This is the important part. The other category is where you discover requests nobody thought to name. Read a sample, add a rule, run again. Two or three passes usually get "other" small.

Источник

The tally CSV. Five examples per category, including one that went badly. A note on which categories need a human decision (prices, refunds, complaints).

Источник