What improvement was reported when using 'please' in prompts?

Matt WolfeFutureTools.io · Founder of Future Tools | Host of Next Wave Podcast

How to Get 10X Better Results from LLMs

13 Nov 2025 · 0 min

Watch on YouTube

0:17

Short YouTube transcript excerpt where two speakers discuss how adding emotions or polite language (e.g., saying “please”) when prompting AI appears to improve results. One speaker cites research claiming about a 7% improvement in answer accuracy when prompts include “please”; the other speaker notes people still say “please” out of habit or hope.

Questions this video answers

Approximately 7% better (more accurate answers).

It sounds silly, but whenever you add these emotions into your text when you talk to AI, it tends to do better job. I don't know why even when you say please. You know, there's even research where they measure the difference between please and without please, and with 7% better, more accurate answers when you have please.
I've heard people try to mute that, but I still say please anyways, because I'm just like, maybe it works.