jake/clawdbot-workspace
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
clawdbot-workspace/proposal-factory/PIPELINE.md

6.9 KiB
Raw Blame History

Proposal Factory — Pipeline Operator Guide

Overview

This system takes a client intake (form OR meeting transcript) and produces a complete proposal package:

  1. Spec Sheet — Technical scope, MCP servers needed, agent team, automation tiers
  2. Visual Deliverables Doc — The "What You Get & What It Does" PDF with AI-generated visuals
  3. Contract/SOW — Legally ready statement of work with payment terms
  4. Discovery Questions — Tailored follow-up questions based on their industry/use case

How to Use

Option A: Intake Form

Jake fills out intake/CLIENT-NAME.md using the template at templates/intake-form.md.

Option B: Meeting Transcript

Jake drops a transcript (text, audio file, or paste) and tells Buba: "New prospect intake from this transcript." Buba extracts the intake fields and creates the intake file.

Trigger

Tell Buba: "Run proposal factory for [CLIENT NAME]"

Buba will:

  1. Read the intake file at intake/CLIENT-NAME.md
  2. Generate all outputs into output/CLIENT-NAME/
  3. Send the package to Jake for review

Pipeline Stages

Stage 1: Parse Intake

  • Input: intake/CLIENT-NAME.md (form) OR raw transcript
  • Action: Extract all fields into structured data
  • Output: Validated intake data (in memory)

Stage 2: Scope & Architecture

  • Input: Validated intake data
  • Action:
    • Map client's tech stack to MCP servers (match against our library of 30+)
    • Map their team roles to agent personas
    • Map their services/pain points to automation capabilities
    • Determine tier (Starter / Standard / White Glove) from scope
    • Calculate pricing using tier logic
  • Output: output/CLIENT-NAME/spec-sheet.md

Stage 3: Visual Deliverables Doc

  • Input: Spec sheet + intake data
  • Action:
    • Generate 4 Nano Banana Pro visuals:
      1. Work OS dashboard mockup (with THEIR platform names)
      2. Factory pipeline (with THEIR deliverable types)
      3. Agent chat interface (with THEIR agent names + data sources)
      4. AI-generated creatives (with THEIR industry examples)
    • Build HTML from template, inject client-specific content
    • Convert to PDF via Puppeteer + Ghostscript
    • Add case studies matched to their use case
  • Output: output/CLIENT-NAME/deliverables.pdf + output/CLIENT-NAME/deliverables.html

Stage 4: Contract/SOW

  • Input: Spec sheet + pricing + client info
  • Action:
    • Select contract template based on tier
    • Fill in: client name, company, scope, integrations list, deliverables, pricing, timeline
    • Generate HTML + PDF
  • Output: output/CLIENT-NAME/contract-sow.pdf + output/CLIENT-NAME/contract-sow.md

Stage 5: Discovery Questions (Optional)

  • Input: Intake data + industry
  • Action: Generate tailored discovery questions based on gaps in intake + industry-specific concerns
  • Output: output/CLIENT-NAME/discovery-questions.md

Stage 6: Package & Deliver

  • Input: All outputs
  • Action: Send summary to Jake with file links, ready for review
  • Output: Message to Jake with the full package

Tier Logic

Starter — $2,499

  • 1 messaging channel
  • 1-3 MCP integrations
  • Basic agent configuration (1 agent)
  • 5-7 day delivery
  • No custom visuals in proposal (text-only spec sheet + contract)

Standard — $7,499

  • 2-3 messaging channels
  • 4-8 MCP integrations
  • 2-3 agents with custom personas
  • 14-day delivery
  • Visual deliverables doc (no case studies)

White Glove — $20,000+

  • 3+ messaging channels
  • 8+ MCP integrations
  • 3+ agents with deep workflow logic
  • Work OS + Factory pipeline
  • 6-week delivery
  • Full visual deliverables doc + case studies + discovery questions
  • Post-launch retainer offered

Custom Enterprise — $20,000-60,000+

  • Multi-phase rollout
  • 5+ agents across business pillars
  • 10+ MCP integrations
  • Custom expansion roadmap
  • Full proposal package + phased pricing

Case Study Matching

Client Use Case Matched Case Study
CRM + lead gen + pipeline *A: LocalBosses (GHL + multi-tool Work OS)
Bulk content/deliverable production *B: MCP Factory (30+ servers, 18-stage pipeline)
Multi-channel agent (Discord/Telegram/iMessage/SMS) *C: Buba (multi-channel coaching agent)
Ad management (Google/Meta) *A: LocalBosses (Google Ads + Meta Ads integration)
Automation/workflow generation *A: LocalBosses (n8n workflow generation via chat)
Client-facing chatbot *C: Buba (24/7 always-on, persistent memory)
Sales enablement / coaching *C: Buba (structured daily coaching, cron scheduling)
Data/analytics dashboards *A: LocalBosses (cross-platform dashboard, iframeable)

Nano Banana Prompt Library

Work OS Dashboard

Professional dark-mode SaaS dashboard mockup on a laptop screen. Dashboard title: '[CLIENT] WORK OS'. Shows unified analytics pulling from [PLATFORM LIST]. Include: cross-platform analytics chart, [INDUSTRY-SPECIFIC METRIC] tracker, pipeline funnel, activity feed. Blue/purple accent colors, modern glassmorphism cards.

Factory Pipeline

Clean product marketing mockup showing an AI Factory pipeline on a monitor. 5 connected stages: Stage 1 'Data Pull' (icons for [THEIR TOOLS]), Stage 2 'AI Generation' ([THEIR DELIVERABLE TYPES]), Stage 3 'Brand Customization' (logos being swapped per client), Stage 4 'Quality Review' (human approval checkpoint), Stage 5 'Delivery' (outputs going to clients). Dark UI with gradient cards. Title: 'THE FACTORY'.

Agent Chat Interface

Product mockup showing AI agent chat on phone and laptop side by side. Phone shows [MESSAGING PLATFORM] chat where user asks '[INDUSTRY-SPECIFIC QUESTION]' and agent [AGENT NAME] responds with embedded dashboard card showing [RELEVANT METRICS FROM THEIR TOOLS]. Laptop shows expanded dashboard. Modern, sleek. Title: 'ASK ANYTHING, GET EVERYTHING'.

AI Creatives

Product mockup showing AI-generated branded marketing content. Show: [NUMBER] different [INDUSTRY] branded [DELIVERABLE TYPES] with different logos and color schemes, social media post graphics for LinkedIn and Facebook, [AD PLATFORM] ad creatives, and a branded [REPORT TYPE] cover. AI sparkle icons. Banner: 'AI-GENERATED CREATIVES IN MINUTES'. Overhead flat-lay, clean background.

File Locations

  • Intake template: templates/intake-form.md
  • Deliverables HTML template: templates/deliverables-template.html
  • Contract templates: templates/contract-starter.md, templates/contract-standard.md, templates/contract-whiteglove.md
  • Discovery questions template: templates/discovery-questions-template.md
  • Case study assets: case-studies/ (images + text blocks)
  • Output: output/CLIENT-NAME/ (one folder per prospect)

Maintenance

  • When we ship a new project: Add it as a case study in case-studies/
  • When pricing changes: Update tier logic in this doc + contract templates
  • When we build new MCP servers: They auto-expand our capability matching