Using Conclave

Run multi-agent deliberation sessions with evidence-based debate and structured decision-making

Conclave is Mega Brain’s multi-agent deliberation system where experts debate strategic decisions with evidence-based reasoning. This guide shows you how to run Conclave sessions and interpret results.

What is Conclave?

Conclave brings together multiple AI agents to deliberate on strategic decisions:

1. Phase 0: Constitutional Foundation

   Invokes core principles (Empiricism, Pareto, Inversion, Antifragility)

2. Phase 1: Role Agent Debate

   C-Level agents (CRO, CFO, CMO, COO) present positions

3. Phase 2: Methodological Critic

   Evaluates quality of reasoning and evidence

4. Phase 3: Devil's Advocate

   Identifies fragile premises and risks

5. Phase 4: Synthesizer

   Integrates perspectives into actionable recommendation

Conclave runs ONCE per query - no loops. If confidence < 60%, it escalates to human decision.

Basic Conclave Session

Invoking Conclave

/conclave "Should I change closer commission from 10% to 15%?"

High Confidence (>60%)

═══════════════════════════════════════════════
           CONCLAVE SESSION
═══════════════════════════════════════════════

QUERY: "Should I change closer commission 
        from 10% to 15%?"

DATE: 2026-03-06
VALUE AT RISK: ~$180K/year

═══════════════════════════════════════════════
PHASE 0: CONSTITUTIONAL FOUNDATION
═══════════════════════════════════════════════

┌──────────────────────────────────────────────┐
│  📜 CONSTITUTION INVOKED                  │
├──────────────────────────────────────────────┤
│                                            │
│  ⚖️ EMPIRICISM                             │
│  Decisions based on DATA, not intuition  │
│  → This deliberation must cite numbers   │
│                                            │
│  📊 PARETO (80/20)                         │
│  Focus on highest leverage actions       │
│  → Which option has most ROI?           │
│                                            │
│  🔄 INVERSION                              │
│  Ask what would make this FAIL           │
│  → Agents must name failure modes       │
│                                            │
│  💪 ANTIFRAGILITY                          │
│  Prefer options that benefit from chaos  │
│  → Which option gets stronger under stress?│
│                                            │
└──────────────────────────────────────────────┘

═══════════════════════════════════════════════
PHASE 1: ROLE AGENT DEBATE
═══════════════════════════════════════════════

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ CRO (Chief Revenue Officer)            ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

"I support the increase, BUT with conditions.

CURRENT DATA:
• Close rate: 58% (below Cole Gordon's 60% benchmark)
• Avg deal size: $12K
• Current comp: $1,200/deal (10%)
• Turnover: 40% annually

INCREASING TO 15% ($1,800/deal) ONLY MAKES SENSE IF:

1. We're losing top performers to competitors
   [Evidence: Exit interviews show 2 left for higher comp]
   
2. Increased retention > cost increase
   Math: If we retain 2 closers, save $60K in training
   
3. We tie it to performance threshold
   Proposal: 15% only if close rate >= 60%
   
RISK: We throw money at symptom (turnover) not 
      cause (poor script/training)

Sources: [CG001_045] Cole Gordon close rate benchmark
         [AH002_023] Unit economics framework

CONFIDENCE: 0.72"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ CFO (Chief Financial Officer)          ┃  
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

"The math says NO, unless structure changes.

UNIT ECONOMICS:
Current: $12K deal × 58% close = $6,960 per opp
         Commission: $1,200 (17.2% of expected value)
         
Proposed: $12K × 58% = $6,960 per opp
          Commission: $1,800 (25.8% of expected value)
          
IMPACT: +$600/deal = $180K annually (30 deals/closer/year)

BREAK-EVEN: We'd need EITHER:
  • Close rate to increase to 87% (unrealistic)
  • Deal size to increase to $18K
  • Volume to increase 50%
  
NONE of these are likely just from raising comp.

ALTERNATIVE:
Test performance-based tiers:
  • Base: 10% (<60% close rate)
  • Tier 1: 12% (60-70% close rate)  
  • Tier 2: 15% (>70% close rate)
  
This creates upside without guaranteed cost increase.

Sources: [AH001_023] LTV:CAC framework

CONFIDENCE: 0.85"

═══════════════════════════════════════════════
PHASE 2: METHODOLOGICAL CRITIC
═══════════════════════════════════════════════

QUALITY SCORE: 74/100

Breakdown:
• Premises declared:      18/20 (both stated assumptions)
• Evidence traceable:     16/20 (good chunk_id usage)
• Logic consistent:       15/20 (CFO math has gap)
• Alternative scenarios:  15/20 (CRO mentioned 1, CFO 1)
• Conflicts resolved:     10/20 (tension not addressed)

CRITICAL GAPS:

1. MISSING DATA: What's industry benchmark for 
   closer compensation?
   
2. CAUSAL ASSUMPTION: Both assume comp → retention
   but no evidence turnover is comp-related
   
3. OPPORTUNITY COST: Neither addressed what else
   $180K could buy (training? better leads?)

RECOMMENDATION: REVISE
Needs: Turnover root cause analysis before deciding

═══════════════════════════════════════════════
PHASE 3: DEVIL'S ADVOCATE
═══════════════════════════════════════════════

MOST FRAGILE PREMISE:
"Turnover is caused by low compensation"

WHY IT'S WEAK:
Exit interviews are notoriously unreliable. People
say "money" when real reason is management, culture,
or lack of growth.

RISK NOT DISCUSSED:
Entitlement Spiral

12 months from now:
"We raised commission to 15% but performance didn't 
improve. Now closers feel ENTITLED to 15% and any 
reduction would cause mass exodus. We're trapped."

Probability: 60%
Impact: $180K/year permanently locked in

SCENARIO OF REGRET:
You raise to 15%. Close rate stays at 58%. Turnover 
continues because real problem was weak onboarding.
Now you've spent $180K and gotten nothing.

ALTERNATIVE IGNORED:
Invest $50K in:
  • Sales training program
  • Better lead qualification
  • Script optimization
  
These address ROOT CAUSE (skill) not symptom (comp).

═══════════════════════════════════════════════
PHASE 4: SYNTHESIS
═══════════════════════════════════════════════

DECISION RECOMMENDED:

DO NOT raise commission to 15% across the board.

INSTEAD, implement 3-phase approach:

PHASE 1 (30 days): Diagnose turnover root cause
  • Conduct proper exit interviews (3rd party)
  • Survey current closers on satisfaction  
  • Benchmark comp vs. market
  • Cost: $5K
  
PHASE 2 (60 days): Performance tier system
  • <60% close rate: 10% commission
  • 60-70% close rate: 12% commission
  • >70% close rate: 15% commission
  • Cost: Variable, performance-based
  
PHASE 3 (90 days): Training investment
  • Script optimization workshop
  • Objection handling bootcamp
  • Cost: $25K
  
TOTAL FIRST-YEAR COST: ~$80K (vs. $180K blanket raise)
EXPECTED ROI: If close rate ↑ to 65%, generates $360K

MODIFICATIONS FROM DEBATE:
• Added Phase 1 (Devil's Advocate point)
• Changed to tiered structure (CFO suggestion)
• Added training component (Critic gap)

CONFIDENCE: 78%

This is high enough to recommend action, but includes
built-in checkpoints to course-correct.

RISKS RESIDUAL:
• Tier system adds complexity: Mitigate with clear docs
• May not solve cultural issues: Monitor in Phase 1
• Training may not transfer: Use role-play validation

NEXT STEPS:
1. Approve Phase 1 diagnostic - Finance: CFO - By: Mar 20
2. Design tier structure draft - Sales: CRO - By: Mar 27  
3. RFP for training vendor - Ops: COO - By: Apr 3

REVERSION CRITERIA:
IF after 90 days turnover ≥ 40% THEN escalate to CEO
IF tier system causes team conflict THEN simplify

═══════════════════════════════════════════════

Low Confidence (<60%)

═══════════════════════════════════════════════
    [CONCLAVE: DECISION INCONCLUSIVE]
═══════════════════════════════════════════════

⚠️ CONFIDENCE: 48% - BELOW THRESHOLD

TYPE OF UNCERTAINTY:
[✓] Data insufficient
[ ] Irresolvable conflict
[ ] Outside knowledge scope

OPTIONS FOR HUMAN DECISION:

OPTION A: Increase to 15% universally
  Trade-off: Higher retention vs. $180K cost
  Defended by: CRO
  Evidence: Exit interviews (weak)
  
OPTION B: Tiered commission structure  
  Trade-off: Complexity vs. performance incentive
  Defended by: CFO
  Evidence: Unit economics (strong)
  
OPTION C: Investigate root cause first
  Trade-off: 30-day delay vs. better decision
  Defended by: Devil's Advocate
  Evidence: Turnover causation unclear

WHAT'S MISSING:
• Actual turnover root cause data
• Industry compensation benchmarks
• Exit interview reliability validation

⚠️ This case requires HUMAN DECISION.
The Council is NOT recommending any option.

RECOMMENDED ACTION:
Gather missing data (Option C) before deciding.

═══════════════════════════════════════════════

Constitutional Principles

Every Conclave session starts by invoking core principles:

The 4 Constitutional Principles

1. Empiricism ⚖️

Principle: Decisions based on DATA, not opinions or intuitions.

Enforcement:

• Agents must cite specific numbers
• Sources must be traceable (chunk_ids)
• Anecdotes must be labeled as such

Violation example:

❌ "I think customers prefer higher prices"
✓ "In 3 surveys (n=240), 67% preferred premium tier [chunk_AH002_034]"

2. Pareto (80/20) 📊

Principle: Focus on the 20% of actions that generate 80% of results.

Enforcement:

• Agents must identify highest-leverage option
• Compare ROI across alternatives
• Avoid “do everything” solutions

Violation example:

❌ "We should improve marketing, sales, ops, and product"
✓ "Focus on sales (60% of impact) before marketing [CG001_045]"

3. Inversion 🔄

Principle: Ask “What would make this FAIL?” before deciding.

Enforcement:

• Devil’s Advocate MUST list failure modes
• Each agent names 1 risk
• Synthesizer addresses top 3 risks

Example:

"What could make the tiered commission system fail?
• Closers game the system
• Adds reporting complexity
• Creates internal competition/resentment"

4. Antifragility 💪

Principle: Prefer options that BENEFIT from volatility and uncertainty.

Enforcement:

• Identify which option improves under stress
• Avoid fragile solutions that require perfect conditions
• Build in optionality

Example:

Option A: Fixed 15% commission
  Fragile: Locks in cost regardless of performance
  
Option B: Performance tiers  
  Antifragile: Rewards high performance, costs scale with results
  
Winner: Option B (benefits from variation in performance)

Understanding the Agents

Role Agents (Phase 1)

C-Level Executives

Bring functional perspectives:

CRO (Chief Revenue Officer)

• Focus: Growth, sales performance, revenue
• Bias: Toward actions that increase deals
• Sources: Sales frameworks, closing techniques

CFO (Chief Financial Officer)

• Focus: Unit economics, profitability, risk
• Bias: Toward conservative financial decisions
• Sources: Financial frameworks, metrics

CMO (Chief Marketing Officer)

• Focus: Positioning, demand generation
• Bias: Toward brand and marketing investment
• Sources: Marketing strategies, funnels

COO (Chief Operations Officer) - Focus: Scalability, efficiency, execution; Bias: Toward operational improvements; Sources: Process frameworks, systems

Conclave Agents (Phases 2-4)

Methodological Critic

Role: Evaluate QUALITY of reasoning, not merit of decision.

Scores (0-100):

• Premises declared (0-20)
• Evidence traceable (0-20)
• Logic consistent (0-20)
• Alternative scenarios (0-20)
• Conflicts resolved (0-20)

Does NOT:

• Take a position
• Recommend an option
• Inject new data

Devil’s Advocate

Role: Identify fragile premises and hidden risks.

Delivers:

• Most fragile premise (with reasoning)
• Main undiscussed risk
• Alternative scenario that was ignored
• Reversal criteria (when to reconsider)

Next Steps

Working with Agents

Learn how to interact with individual agents

Session Management

Save and resume Conclave sessions