The LLM Council Prompt: Three Adversarial Voices, One Decision
Run this single prompt and get three distinct adversarial AI perspectives on any decision, stress-testing it before you commit.
Free prompt pack · copy-paste ready · built by the BrainVaultAI team
Most founders use Claude as a yes-machine. You describe your idea, it finds reasons it works. That's dangerous for decisions that actually matter. The Council prompt forces Claude to roleplay three distinct analysts, a bull, a bear, and a systems thinker, then synthesize a final recommendation that accounts for all three views. Use it for: pricing decisions, hiring calls, pivots, build-vs-buy, any choice where your own bias is the biggest risk. Takes 2 minutes. Saves you from expensive conviction errors.
LLM Council Prompt
You are running an internal decision council. Three analysts will each give their honest take, then you will synthesize. Decision to evaluate: [DESCRIBE YOUR DECISION IN 2-3 SENTENCES] Analyst 1, The Bull:
Argue the strongest case FOR this decision. What upside is real? What risks are overstated? What would have to be true for this to work brilliantly? Analyst 2, The Bear:
Argue the strongest case AGAINST. What is the founder/operator blind spot here? What fails first? What is the actual downside scenario, not the sanitized one? Analyst 3, The Systems Thinker:
Ignore the bull/bear framing. What are the second and third-order effects? What does this decision change that the other two missed? What feedback loops does it create? Synthesis:
Given all three perspectives, what is the actual recommendation? State it as a clear directive, not a hedge. If you need more information before recommending, state exactly what information and why. Do not soften any of the analyst takes. I need the real arguments, not diplomatic ones.
Council Debrief Synthesis Prompt
After running the three-voice council on your decision, paste this to synthesize:
"You just gave me three adversarial perspectives on [DECISION]. Now act as a neutral fourth voice: the integrator.
Do the following:
1. Identify the one objection all three voices agree on, even if they framed it differently.
2. Identify the one assumption all three voices accepted without questioning it.
3. Identify the strongest argument from each voice and rank them by how much they should actually change my decision.
4. Give me the decision you would make if you had to commit right now, with a one-sentence justification.
5. Tell me what information, if I had it, would flip your recommendation.
Do not summarize what each voice said. Synthesize. I need a decision, not a recap."
The council generates pressure. This prompt converts that pressure into a commit-ready answer.
Want the editable version?
We'll send the editable version of this asset, plus a short series of our highest-leverage AI plays for operators. No spam. Unsubscribe in one click.
Free. No spam. Unsubscribe in one click. By submitting you agree to receive emails from BrainVaultAI.