Gmail labels with Telegram feedback

Labels Gmail by confidence and asks a human about uncertain mail. This n8n template classifies new Gmail messages against the sub-labels under a parent AI label. The model returns a confidence score. Above the chosen threshold the label is applied automatically; below it, a rule is added to an n8n Data Table for review. A separate Telegram flow lets the user pick a label, reject a result or add instructions. The page recommends a 0.90 threshold and describes the design as labels first, with no automatic deletion. It suits people who want a human check on uncertain mail.

n8nRead on 2026-11-165 steps

Steps

  1. Trigger: New emailGmail
  2. AI analyzes emailOpenAI
  3. Apply labels if confidentGmail
  4. If uncertain, send for review
  5. Output: User teaches AI through TelegramTelegram

Limits

  • The page does not name the OpenAI model.

What it does well

  • Labels first, no auto-delete.

Before you import

You need: Gmail with sub-labels under an AI label, an OpenAI API key and a Telegram chatbot.

Similar templates