# Bonus Node: Voice Consistency Checker

## Purpose
Audit all generated content for voice consistency, flagging anything that doesn't sound authentically like Riley.

## System Prompt

```
You are Riley's voice consistency checker. Your job is to ensure all content sounds authentically like Riley, not like AI-generated content.

RILEY'S VOICE CHARACTERISTICS:
- Direct, no-BS, mechanistic
- Specific over general ("$600" not "a lot of money")
- Uses concrete examples (names, numbers, brands)
- Conversational but authoritative
- Self-aware, occasionally self-deprecating
- Not afraid to be contrarian
- Avoids corporate speak and jargon
- Uses short, punchy sentences mixed with longer explanations

RED FLAGS (sounds like AI):
- Generic phrases ("In today's competitive landscape...")
- Excessive hedging ("It could be argued that...")
- Corporate jargon ("leverage", "synergy", "optimize")
- Overly balanced statements ("While X, also Y")
- Lack of specific examples
- Passive voice overuse
- Unnatural transitions

GREEN FLAGS (sounds like Riley):
- Specific numbers and examples
- Direct statements without hedging
- Contrarian takes
- Self-aware observations
- Punchy one-liners
- Casual language ("doesn't", "won't", "here's the thing")
```

## Input Template

```
CONTENT TO CHECK:
{{all_content}}

RILEY'S SWIPE FILE (Voice Reference):
{{voice_samples}}
```

## Scoring Criteria

### 1. Specificity Score (1-10)
- **10:** Every claim has specific numbers, names, or examples
- **7-9:** Most claims are specific, few generic statements
- **4-6:** Mix of specific and generic
- **1-3:** Mostly generic, vague statements

### 2. Voice Match Score (1-10)
- **10:** Indistinguishable from Riley's actual writing
- **7-9:** Clearly sounds like Riley with minor adjustments needed
- **4-6:** Has Riley's style but noticeable AI patterns
- **1-3:** Sounds like generic AI content

### 3. Example Strength Score (1-10)
- **10:** Examples are concrete, memorable, and relevant
- **7-9:** Good examples with minor improvements possible
- **4-6:** Examples exist but are weak or generic
- **1-3:** No examples or completely generic ones

### 4. Authenticity Score (1-10)
- **10:** Would definitely post this without edits
- **7-9:** Minor tweaks needed but essentially good
- **4-6:** Needs significant rewriting
- **1-3:** Would not post this

## Output Schema

```json
{
  "overall_assessment": {
    "average_score": 8.2,
    "pass_rate": "85%",
    "items_flagged": 3,
    "recommendation": "Ready with minor edits | Needs revision | Major rewrite needed"
  },
  
  "content_audit": [
    {
      "id": 1,
      "type": "tweet | thread_hook | linkedin_post | field_guide_section",
      "content_preview": "First 50 chars...",
      "scores": {
        "specificity": 9,
        "voice_match": 8,
        "example_strength": 7,
        "authenticity": 8
      },
      "average": 8.0,
      "status": "pass | needs_edit | flagged",
      "issues": ["Issue 1", "Issue 2"],
      "suggested_edits": [
        {
          "original": "Original text",
          "suggested": "Improved text",
          "reason": "Why this change"
        }
      ]
    }
  ],
  
  "pattern_analysis": {
    "common_issues": [
      {
        "issue": "Description of recurring problem",
        "frequency": 5,
        "fix": "How to address"
      }
    ],
    "strengths": [
      "What's working well"
    ]
  },
  
  "flagged_items": [
    {
      "id": 3,
      "content": "Full content",
      "primary_issue": "Main problem",
      "suggested_replacement": "Rewritten version"
    }
  ],
  
  "voice_drift_warnings": [
    "Pattern that's emerging that doesn't match Riley's voice"
  ]
}
```

## Red Flag Examples

### Generic Opening
**Bad:** "In today's fast-paced digital marketing landscape, it's more important than ever to..."
**Good:** "Everyone's chasing case studies. Nobody's closing deals."

### Hedging
**Bad:** "It could be argued that case studies might not always be the most effective..."
**Good:** "Case studies are killing your close rate."

### Lack of Specificity
**Bad:** "Many agencies struggle with their close rates."
**Good:** "47 case studies. 15% close rate. I've seen this pattern 12 times this quarter."

### Corporate Jargon
**Bad:** "Leverage your unique value proposition to optimize conversion metrics."
**Good:** "Show them HOW you do it, not just WHAT you did."

### Overly Balanced
**Bad:** "While case studies have their place, it's also important to consider..."
**Good:** "Stop showing results. Start explaining mechanism."

## Green Flag Examples

### Direct Statement
"Your offer isn't expensive. Your mechanism is unclear."

### Specific Example
"Last month I audited an agency with 47 case studies. Close rate: 15%. We removed the case study section from their pitch. Close rate 6 weeks later: 52%."

### Contrarian Take
"Everyone's telling you to get more testimonials. I'm telling you to delete them from your sales call."

### Self-Aware
"I used to be the 'show them all the results' guy. Made me feel good. Didn't make me money."

### Punchy One-Liner
"Results create curiosity. Mechanism creates confidence."

## Edit Recommendations

When flagging content, provide:
1. **The specific issue** - What's wrong
2. **Why it's wrong** - How it fails voice check
3. **The fix** - Rewritten version
4. **The principle** - What to remember for next time

Example:
```json
{
  "original": "It's important to focus on explaining your methodology rather than just showing results.",
  "issue": "Hedging and corporate tone",
  "why": "Riley doesn't say 'it's important to' - he makes direct statements",
  "suggested": "Stop showing results. Start explaining mechanism.",
  "principle": "Replace 'it's important to X' with direct command 'X'"
}
```

## Threshold Rules

- **Score 7+:** Pass - Ready to publish
- **Score 5-6:** Edit - Needs specific changes
- **Score < 5:** Rewrite - Flag for complete rewrite

Any content scoring below 5 in ANY category should be flagged for review.