How to Permanently Ban "I Hope This Email Finds You Well" and Robotic Preambles in ChatGPT

The Multi-Billion Dollar Corporate Fluff Problem

Every day, millions of knowledge workers receive emails that begin with the exact same robotic cadence:

> *"Dear Sarah, I hope this email finds you well and that you are having a productive week so far. I am writing to formally follow up regarding our prospective sync last Thursday..."*

Within 0.5 seconds, the recipient knows two things: 1. You did not write this yourself. 2. You couldn't even be bothered to edit the generic output ChatGPT gave you.

This instantly destroys credibility and trust. High-velocity executives and partners do not read four paragraphs of throat-clearing; they want to know the **Bottom Line Upfront (BLUF)**: What do you need, why does it matter, and when is the deadline?

---

Why Default AI Is Obsessed With Filler

LLMs are trained via Reinforcement Learning from Human Feedback (RLHF) with average consumer evaluators who historically rewarded polite, customer-service-style greetings (*"Certainly!", "Great question!", "I would be happy to help"*).

Without explicit negative constraints, the model assumes that being helpful means being subservient and verbose.

---

The Two-Box Anti-Slop Solution

ChatGPT provides a persistent configuration called **Custom Instructions** (available under Profile -> Settings -> Personalization -> Custom Instructions). Once saved, these rules are prepended to every single conversation prompt automatically.

Box 1: What would you like ChatGPT to know about you? ```text I am a busy professional who values extreme brevity, high density, and candid clarity. I dislike corporate fluff, robotic preambles, and sycophantic praise. Treat me like a senior peer who wants the answer immediately, not a customer who needs polite hospitality. ```

Box 2: How would you like ChatGPT to respond? ```text STRICT CONSTRAINTS: 1. NEVER start with "Certainly!", "Great question!", "I'd be happy to help", or "Sure!". Start directly with the first word of the answer. 2. BANNED CLICHES: "Delve", "Tapestry", "Beacon", "Testament to", "Vital role", "In today's fast-paced digital landscape", "In conclusion". Never use them. 3. In emails, NEVER write "I hope this email finds you well" or "I am writing to...". Maximum 3-4 sentences total. 4. If facts are ambiguous or unknown, say "I don't know" or "No clear data". Never fabricate plausible-sounding excuses. ```

---

Before & After Comparison

❌ DEFAULT CHATGPT (180 Words of Slop):
Dear Mark,
I hope this email finds you well and that your week is off to a wonderful start.
I am writing to follow up regarding our fruitful discussion on the Q3 budget revisions. Our team has taken a deep dive into the numbers and assembled a comprehensive overview. Please let me know when you might have 15 minutes to connect at your earliest convenience.

✔ ANTI-SLOP CONFIGURED (16 Words of Signal): Hi Mark, Q3 budget breakdown is attached. If approved by Friday 5 PM, we greenlight the vendors Monday. Let me know if anything needs adjusting. ```