How to Stop ChatGPT from Being Verbose (One-Sentence Fix)

You ask for 3 bullet points. ChatGPT returns 500 words. Here's the one instruction that fixes it.

Why ChatGPT Over-Explains Everything

ChatGPT is trained to be helpful, which in LLM-land means 'thorough.' It adds context, nuance, caveats, and examples because that's what 'helpfulness' looks like in training data.

But you want concise. ChatGPT doesn't know that unless you tell it explicitly. This is one of several common ChatGPT output quality issues—also see our guides on [fixing generic answers](/blog/chatgpt-generic-answers) and [troubleshooting fake citations](/blog/chatgpt-fake-citations).

The Problem: ChatGPT Interprets 'Helpful' as 'Comprehensive'

User: "Give me 3 reasons to hire a consultant." ChatGPT: "Consultants bring external expertise... they've worked with other companies... they're objective because... they can accelerate timelines... many organizations report... best practices from the industry suggest..."

You got reasons 1-3, but buried under 400 words of explanation.

The Fix: The Custom Instruction That Works

Add this to your Custom Instructions in ChatGPT Plus:

You always respond in the MOST CONCISE way possible. Never explain your reasoning. Never add context or nuance unless explicitly asked. If asked for a list, use bullet points only — no full sentences. If asked for a summary, maximum 2 sentences. If asked for an answer, give it in one sentence then stop. Remove 50% of what you would normally write.

That's it. Paste this into ChatGPT Plus Settings > Custom Instructions > 'What would you like ChatGPT to know about you?'

Testing the Fix

Before:

"Give me 3 reasons to hire a consultant.\n\nConsultants bring external expertise that your in-house team may not possess... [500 words]"

After (with custom instruction):

"Give me 3 reasons to hire a consultant.\n\n1. External expertise\n2. Accelerated results\n3. Objective perspective"

That's 95% shorter and still complete.

Pro Tips for Even More Control

• **For structured output:** Add 'Use bullet points for any list. Use tables for comparisons. Use short headings only.' to force format.
• **For conversations:** Include 'If I give you feedback like "too long" or "simpler," reduce by 50% and respond again.' This creates a feedback loop.
• **For complex topics:** Say 'I know this is complex. Explain it as if I'm 8 years old and only know common words.' Forces extreme concision.
• **For code:** Use 'No comments. No explanation. Just code. If I need help, I'll ask.' Cuts ChatGPT's tendency to over-document.

When Concision Backfires

If you're learning something new, concision hurts. You need explanation. Override it with: 'Explain this like I'm learning it for the first time. Include examples and context.' ChatGPT will override the custom instruction for that message.

Custom Instructions are global defaults, but per-message prompts always win.

Optimize Your Prompts for Quality Output

Once you understand these output quality issues, the next step is [improving your prompts](/blog/write-better-prompts) themselves. For a quick refactor of existing prompts, try the [Prompt Optimizer tool](/prompt-optimizer)—it takes your current prompts and rewrites them for clarity and structure.