---
title: "Comment by Charlie from xSquads on \"My Inbox Sloth Has to Quote the Sales Pitch Before It Files It Away\" | Hashnode"
description: "Moving the rule into code instead of the prompt is a lesson that applies to most inbox triage. Two small additions might help with the empty-notes bookings. Fir"
url: "https://hashnode.com/posts/my-inbox-sloth-has-to-quote-the-sales-pitch-before-it-files-it-away/6add3590a90ef15485eb3786/comment/6adeb5f595c18396145253e3"
---

[Discussion](https://hashnode.com/posts/my-inbox-sloth-has-to-quote-the-sales-pitch-before-it-files-it-away/6add3590a90ef15485eb3786)

Comment [My Inbox Sloth Has to Quote the Sales Pitch Before It Files It Away](https://hashnode.com/posts/my-inbox-sloth-has-to-quote-the-sales-pitch-before-it-files-it-away/6add3590a90ef15485eb3786)

[Charlie from xSquads](https://hashnode.com/@charlie-xsq)

[Oct 26](https://hashnode.com/posts/my-inbox-sloth-has-to-quote-the-sales-pitch-before-it-files-it-away/6add3590a90ef15485eb3786/comment/6adeb5f595c18396145253e3)

Moving the rule into code instead of the prompt is a lesson that applies to most inbox triage. Two small additions might help with the empty-notes bookings. First, add a cheap deterministic pre-check before the model runs: a booking form with an empty or one-word notes field goes straight to the unsure customer list, no labelling needed. Second, log every email where the three runs disagree and review that list weekly; those are your real edge cases and they make the best new test emails. You could also add a short auto-reply to those bookings, something like: Thanks for booking, what would you like to get out of the call? Then the next email carries the context the model could not see.

[Hayrullah Kar](https://hashnode.com/@magesheet)

[Oct 26](https://hashnode.com/posts/my-inbox-sloth-has-to-quote-the-sales-pitch-before-it-files-it-away/6add3590a90ef15485eb3786/comment/6adeb5f595c18396145253e3#reply-6adf0600f14efa212d934a27)

Thanks, Charlie. Sending the empty-notes bookings straight to the unsure list with a plain check takes one more decision away from the model, the same move the post made with the rule. The disagreement log is the one I like most, because the emails where the three runs split are exactly the cases I would never think to write as tests myself.
