Tier 2: Flexible Templates

Meta-Template Extraction Guide

Universal framework for generating 10 template types with speed and flexibility

🔧 Framework-Based Approach

📖 Quick Navigation

Philosophy & Goals
3-Layer Architecture
10 Template Types
[REPLACE:] Token Convention
LLM Build Protocol
Shared Design DNA
Validation Rules
vs Playbook Blueprint

Philosophy & Goals

Use Meta-Template when: You need to generate high-volume, flexible templates across different document types. Email sequences, assessments, reports, dashboards, newsletters, content deliverables.

Core principle: One extraction framework → N template types. Reusable architecture. Speed over depth.

When Tier 2 Wins

Email sequences (5-7 emails in a campaign)
Assessment workflows (scoring logic + results)
Reports & dashboards (data visualization templates)
Content libraries (10+ articles, blog posts)
Rapid prototyping (test multiple template variations)

3-Layer Architecture

Layer 1: Metadata (YAML)

Structured data about template purpose, fields, dependencies.

---
template_id: email-sequence-campaign
type: email
category: email-sequences
version: "1.0"
target_audience: prospects
fields:
  - name: prospect_name
    type: text
  - name: value_prop
    type: textarea
  - name: cta_url
    type: url
---

Layer 2: Instructions (Directives)

LLM directives. How to extract, transform, validate content for this template type.

Extract from conversation:
- prospect_pain_points (3-5 key problems)
- our_unique_positioning (1 sentence)
- value_prop_for_prospect (specific outcome)
- next_step_cta (action after email)

Generate 5 emails:
1. Cold hook (attention grab)
2. Problem validation (show you understand)
3. Social proof (case study brief)
4. Solution detail (what we do)
5. CTA & urgency (next step)

Layer 3: Render (HTML/CSS/JS)

Visual output. Layout, typography, interactive elements.

💡 Key insight: Layers are independent. Update Layer 2 (logic) without touching Layer 3 (design). Update Layer 3 (styling) without rewriting Layer 2.

10 Template Types Taxonomy

Type	Purpose	Key Fields	Output Format
Brand Strategy	Positioning docs, brand books	voice_dna, visual_system, messaging_pillars	PDF/DOCX
Workflow	SOPs, process docs, playbooks	steps, dependencies, ownership	PDF/DOCX/HTML
Content Delivery	Blogs, articles, newsletters	headline, outline, cta_url	HTML/Markdown
Sales/Conversion	Email sequences, landing pages	hook, pain_points, cta	HTML/Email
Assessment	Quizzes, scoring rubrics, diagnostics	questions, scoring_logic, results	HTML/Interactive
Expert Extraction	Framework distillation, IP capture	frameworks, principles, mental_models	PDF/DOCX
Email/Sequence	Drip campaigns, follow-ups	email_1_hook, email_2_proof, cta	Email/HTML
Dashboard/Report	Data summaries, analytics	metrics, sparklines, kpis	HTML/PDF
Documentation	API docs, user guides, onboarding	sections, code_examples, faq	HTML/Markdown
Client Deliverable	Custom reports, proposals, analyses	client_name, findings, recommendations	PDF/DOCX

[REPLACE:] Token Convention

Structured placeholders instead of bare tokens. Each token includes context about what it is and where it comes from.

Format: [REPLACE: token_name | context_hint]

Example:

[REPLACE: prospect_company | extracted from "tell me about your business"]
[REPLACE: pain_points | list of 3-5 problems mentioned]
[REPLACE: value_prop | 1-sentence summary of what we do for them]
[REPLACE: cta_url | the link they should click]

Why structured tokens: AI can see context. Reduces hallucination. Clear what each field should contain. Easier to validate during rendering.

LLM Build Protocol

6-Step Build Sequence

Parse template metadata: Understand fields, format, purpose
Extract from source: Pull raw data from conversation/document
Transform to spec: Match data to template fields using [REPLACE:] tokens
Validate completeness: Ensure all required fields populated
Render output: Generate final document (HTML/PDF/Email)
Quality check: Verify formatting, tone, completeness

Shared Design DNA

All Tier 2 templates inherit base design system. CSS variables for consistency.

:root {
  --primary: #3b82f6;
  --secondary: #1e40af;
  --accent: #60a5fa;
  --background: #f9fafb;
  --text-primary: #1a1a1a;
  --text-secondary: #555;
  --border: #e5e7eb;
  --success: #22c55e;
  --warning: #f59e0b;
  --error: #ef4444;
}

Benefit: Update base design once → all templates reflect change instantly.

Validation Rules

✓ DO

Use [REPLACE:] tokens consistently
Include context hints in each token
Validate all required fields populated
Test template with 3+ real data samples
Document metadata clearly

✗ DON'T

Bare tokens like [NAME] or [EMAIL]
Hidden dependencies between fields
One-off custom variations per use
Skip validation step
Mix Layer 2 logic with Layer 3 styling

Meta-Template vs Playbook Blueprint

Dimension	Meta-Template (Tier 2)	Playbook Blueprint (Tier 1)
Best for	High-volume, flexible, reusable	Strategic, persona-specific, deep
Use cases	Email, reports, assessments, content	7-step playbooks, strategic plans
Extraction depth	Shallow (field mapping)	Deep (7 analysis fields)
Output variability	High (10 template types)	Low (fixed 7-step structure)
Persona compounding	Minimal	Maximal (builds moat)
Build time per output	Fast (< 1 min)	Slow (2-5 mins)

Hybrid strategy: Use Tier 2 for volume. Use Tier 1 for strategic moat-building. Together = complete template ecosystem.