---
title: "What Google Just Revealed About AI Prompting (That Works)"
description: "5 source-backed AI claims from What Google Just Revealed About AI Prompting (That Works), with key statistics, context, and the original AI Adopters Club…"
url: "https://kbanc.com/claims-library/what-google-just-revealed-about-ai-prompting-that-works"
source: "https://aiadopters.club/p/what-google-just-revealed-about-ai"
date: "2025-04-13"
topics: ["strategy", "implementation", "tools"]
generated: "2026-08-31"
---

# What Google Just Revealed About AI Prompting (That Works)

By Kamil Banc | April 13, 2025

## Claims

1. **Whitepaper Confirms Clear Communication** (source summary): Google's 68-page prompt engineering whitepaper confirms that clear communication beats technical expertise when writing prompts.
2. **Start With Action Verbs** (source summary): Starting prompts with clear action verbs like Analyze or Compare dramatically improves AI results.
3. **Specify Length and Format** (source summary): Specifying exact length and format, like two paragraphs, produces better AI output than vague requests.
4. **Break Tasks Into Steps** (source summary): AI performs significantly better when complex tasks are broken into smaller, sequential prompts.
5. **Save and Reuse Prompts** (source summary): Keeping a prompt journal and tools like Right-Click Prompt help save and reuse winning prompts.

## Evidence

### Quote
> "The companies getting real value from AI aren't the ones with the most technical knowledge - they're the ones willing to experiment systematically." - Kamil Banc

### Key Statistics
- **68 pages**: Length of Google's prompt engineering whitepaper by Lee Boonstra that the author analyzed for the article
- **5 techniques**: Number of beginner-friendly prompting techniques the author extracted from Google's whitepaper
- **3 misconceptions**: Common prompting myths debunked: needing technical expertise, using more words, and hunting for a perfect prompt
- **3 columns**: Recommended prompt journal format: Prompt, What it did well, What could be better

## Context
The article is a practitioner's distillation of Google's 68-page prompt engineering whitepaper by Lee Boonstra, filtered through Kamil Banc's hands-on experience training business users. Rather than covering advanced methods like token sampling or self-consistency, the author selects five techniques that deliver results immediately: showing examples, leading with action verbs, specifying length and format, decomposing complex tasks, and journaling successful prompts. The methodology emphasizes systematic experimentation over searching for perfect templates, with a simple three-column journal to capture what works. For practitioners, the recommended starting point is to pick one business process, apply the techniques, document outcomes, and iterate, supported by browser-based tools like Right-Click Prompt for storing and reusing winning prompts across ChatGPT, Claude, and other platforms.

## Source
- Original: [What Google Just Revealed About AI Prompting (That Works)](https://aiadopters.club/p/what-google-just-revealed-about-ai)
- Cite: kbanc.com/claims-library/what-google-just-revealed-about-ai-prompting-that-works

## Primary Evidence
- [Right-Click Prompt](https://chromewebstore.google.com/detail/rcp-right-click-prompt-ai/pdhaalepogdcfdkbhgphjppjcmjbjfkd) (chromewebstore.google.com; supports claim 5)