The Trend I’m Watching: “AI Checking AI” and How to Use It Today
by Sandy Waggett
4 min reading time
The Trend I’m Watching: “AI Checking AI” and How to Use It Today
TL;DR: The next big advantage in AI won’t be having the best model but having a process that forces AI to be accurate. By running a “critique pass” on AI outputs—essentially having AI check AI—you can identify hidden assumptions, risks, and improve decision-making. This simple 3-step method takes just 5 minutes and can save your business costly mistakes.
Why “AI Checking AI” Is the Next Big Shift
When it comes to artificial intelligence, many businesses focus on finding the “best AI” with the most advanced capabilities. But the reality is shifting. Leading platforms like Microsoft have already started integrating multiple AI models working together — one to generate content, and another to critique or verify it. Microsoft's recent Copilot multi-model intelligence is a prime example of this trend.
This isn’t just a fancy gimmick; it’s a strategic move toward improving AI reliability and accuracy. The takeaway for business owners and marketers is clear: stop trusting the first AI answer blindly. Instead, build a process that forces AI to self-check and deliver more reliable insights.
How to Implement AI Checking AI in 3 Simple Steps
This approach is straightforward and takes about 5 minutes per important output. Here’s how you can do it today:
Step 1: Get Your First Draft — The Fast Answer
Ask your AI tool (like ChatGPT) for the initial draft. This could be a plan, email, pricing recommendation, strategy memo, or any other business-related content you need quickly. This first pass captures the core idea fast.
Step 2: Run the “Critique Pass” — Pressure-Test the Output
Right after receiving the first answer, copy and paste it again and prompt your AI with a critique role. For example, input this prompt:
You are my AI critic and risk analyst.
Your job is to pressure-test the answer above.
Do this:
- List the top 10 assumptions it makes
- Identify what’s missing that would change the recommendation
- Flag any places it may be hallucinating or overconfident
- Give me the 5 biggest risks if I implement this as-is
- Rewrite the answer to be more accurate, more cautious, and more actionable
Rules:
- If you don’t know something, say “unknown”
- Don’t invent facts
- Keep it scannable
This step is crucial and often skipped but it’s the secret sauce that transforms AI from a one-and-done draft machine into a reliable business partner.
Step 3 (Optional): Run a “Council Pass” — Compare Multiple AIs
If you have access to multiple AI tools (for example, ChatGPT, Google’s Gemini, Anthropic’s Claude), run the same prompt through each one and collect their answers. Then ask one AI to create a “council summary” that:
Highlights where the AIs agree
Points out where they disagree
Evaluates which answer is more credible and why
Combines the best elements into a final, improved answer
This multi-model comparison further reduces risk and boosts confidence in the AI’s recommendations.
Real-World Example: Pricing Strategy Using AI Checking AI
Imagine you want to adjust your pricing structure. Here’s how you can apply this process:
First draft: Ask AI to propose a new pricing model based on your product, market, and goals.
Critique pass: Request the AI critic to list assumptions like time estimates, scope creep, and overhead costs that might impact profitability.
Refinement: Get an improved pricing recommendation that accounts for these factors plus a script to communicate the change clearly to customers.
This simple loop can prevent expensive “oops” moments that happen when pricing changes are implemented without fully understanding hidden risks.
Why This Matters for Your Business
At MSW Interactive Designs, we don’t treat AI as a one-off content generator. Instead, we embed this critique loop into our workflows whenever AI outputs impact strategy, automation, dashboards, or offers. This approach gives us two key advantages:
Speed: AI generates ideas quickly.
Accuracy: The critique pass checks and improves those ideas immediately.
The combination makes AI genuinely usable in business decision-making rather than just a flashy tool.
How You Can Start Today
Don’t wait for advanced multi-model AIs to become standard. You can create your own “AI checking AI” process right now with the tools you already have. Make it a habit to:
Never accept the first AI answer as gospel
Always run a critique pass to pressure-test outputs
Compare multiple AI models if possible to get diverse perspectives
This practice will help you avoid costly mistakes, improve your decision quality, and harness AI’s power responsibly.
About the Author: Sandy Waggett is the founder of MSW Interactive Designs and Infinite Yous, a leading AI integration expert who helps businesses leverage AI for growth.
It's free of charge but insanely valuable. We'll use the time the way you want to. Website or marketing review, AI integration ideas, you name it. Let's start a conversation - we're here to help you be successful online!
