Brainforge Knowledge

GTM_AGENTS_ROADMAP

8 min read

GTM Agents Roadmap

Purpose: Track all GTM tools, processes, and agents built for Brainforge Status: Active Development Last Updated: 2025-01-16

📋 Table of Contents

🎯 Recap: What We’ve Built

Session Overview (January 2025)

Goal: Build GTM agents to replicate Brainforge’s unique BDR tactics and outreach approach, focusing on go-to-market work only (not PM, content, ops).

Approach: Human-in-the-loop agents that draft content for review, prioritizing quality over automation.

🤖 Current Agents

1. Event Follow-Up Agent ✅

Status: Complete & Ready for Testing
Location: gtm/agents/.claude/agents/event-follow-up-agent.md
Playbook: gtm/agents/memory/event-follow-up-playbook.md

What It Does:

  • Generates hyper-personalized event follow-up email sequences
  • Parses event data (attendee lists, booth scans, meeting logs)
  • Segments contacts by engagement type (Booth Visitor, Panelist, VIP Dinner, Meeting Logs, General Attendee)
  • Identifies persona (Executive, Practitioner, VC/Investor) and adjusts tone/messaging
  • Generates multi-touch follow-up sequences (5 touches)
  • Qualifies leads through natural conversation questions
  • Creates detailed handoff summaries for AEs

Key Features:

  • Hyper-personalization with specific event context and engagement details
  • Persona-matched tone (Executives = concise/ROI-driven, Practitioners = tactical/use-case focused, VCs = vision-led)
  • Qualification scoring (1-10 per dimension, max 50 total)
  • Handoff summaries for qualified leads routing to AEs

Model: Based on 11x’s Alice event follow-up system (human-in-the-loop version)

Documentation:

  • Agent Definition: gtm/agents/.claude/agents/event-follow-up-agent.md
  • Playbook: gtm/agents/memory/event-follow-up-playbook.md
  • Example Workflow: gtm/agents/examples/event-follow-up-workflow.md
  • Handoff Ticket: gtm/agents/TICKET_EVENT_FOLLOW_UP_AGENT.md

Next Steps:

  • Test with real event data
  • Collect feedback from team member testing
  • Refine templates based on results
  • Add more case studies/resources

2. LinkedIn Sequence Agent ✅

Status: Complete
Location: gtm/agents/.claude/agents/linkedin-sequence-agent.md
Playbook: gtm/agents/memory/linkedin-playbooks.md

What It Does:

  • Generates multi-step LinkedIn outreach sequences
  • Provides multiple options for each step in the sequence
  • Formats output for human review and GSheets copy-paste
  • Follows established playbooks (Job Posting, Mutual Intro, etc.)

Playbooks Included:

  1. Job Posting to Lead (5-step sequence)

    • I: Initial Connection Request
    • II: Intro Blurb (Follow-up)
    • III: Job Application (Optional)
    • IV: Resource Alignment (CPG/SaaS/General options)
    • V: Thank You / Next Connection

  2. Mutual Intro (Single-step, best performing playbook)

    • 4 blurb options: General Data+AI, Product Analytics, AI Growth, Lifecycle Marketing
    • Selection logic based on prospect focus
    • Highly personalized (e.g., “Honey Stinger’s lifecycle marketing”)

Key Features:

  • Multi-sequence playbooks with multiple options per step
  • Selection logic (when to use each option)
  • Timing and triggers for each step
  • Personalization placeholders and guidance
  • Easy copy-paste into GSheets template

Documentation:

  • Agent Definition: gtm/agents/.claude/agents/linkedin-sequence-agent.md
  • Playbooks: gtm/agents/memory/linkedin-playbooks.md

Next Steps:

  • Add more playbooks (Mutual Intro variations, Vertical Entry, etc.)
  • Test with real prospects
  • Refine selection logic based on performance

3. ICP Analysis Agent (Framework) ✅

Status: Complete (Framework)
Location: gtm/agents/examples/icp-analysis-example.md
Prompt: gtm/agents/prompts/icp-agent-prompt.md

What It Does:

  • Analyzes prospect qualification using ICP framework
  • Uses “thinking slow, thinking fast” methodology (reasoning first, conclusion last)
  • Applies qualification criteria (ICP, BANT, Brainforge-specific questions)
  • Provides structured analysis with scoring and recommendations

Key Features:

  • “Thinking Slow, Thinking Fast” approach (System 2 reasoning before System 1 conclusion)
  • Reasoning-first analysis structure
  • ICP qualification framework
  • Vertical-specific application (e.g., home services)

Documentation:

  • Example: gtm/agents/examples/icp-analysis-example.md
  • Prompt Template: gtm/agents/prompts/icp-agent-prompt.md

Next Steps:

  • Deploy to Slack (see Slack Deployment Guide)
  • Build full ICP Agent with agent definition
  • Add more vertical-specific qualification criteria

📚 Knowledge Base

Core Knowledge Files

  1. ICP & Qualification

    • gtm/agents/memory/ideal-customer.md - ICP profiles (Startups, Scale Ups, Enterprise, Seasonal Champions)
    • gtm/agents/memory/qualification-criteria.md - BANT framework, Brainforge-specific qualification, scoring
    • gtm/agents/memory/vertical-testing.md - Framework for applying ICP to specific industries (home services example)

  2. Messaging & Positioning

    • gtm/agents/memory/positioning.md - How to describe Brainforge, value propositions, messaging by audience
    • gtm/agents/memory/message-templates.md - Email and LinkedIn message templates (structure ready, needs content)

  3. Tactics & Processes

    • gtm/agents/memory/bdr-tactics.md - BDR outreach and prospecting process (structure ready, needs content)
    • gtm/agents/memory/event-follow-up-playbook.md - Event follow-up playbook with segmentation, personas, sequences
    • gtm/agents/memory/linkedin-playbooks.md - LinkedIn playbook structure with multi-sequence options

Knowledge Gaps (To Fill)

  • BDR Tactics (bdr-tactics.md) - Need actual Brainforge BDR process documented
  • Message Templates (message-templates.md) - Need actual email/LinkedIn templates
  • Positioning (positioning.md) - Need Brainforge’s actual positioning and messaging
  • Client Pain Points - Need specific pain points from existing clients (mentioned for vertical-testing.md)

🚀 Roadmap: What’s Next

Phase 1: Knowledge Capture (In Progress)

Goal: Fill in knowledge base with actual Brainforge content

Tasks:

  • Document actual BDR tactics and outreach process
  • Add real email/LinkedIn message templates
  • Document Brainforge’s positioning and messaging
  • Add client pain points and case studies
  • Populate successful outreach examples

Priority: High (agents need this knowledge to work well)

Phase 2: Agent Testing & Refinement

Goal: Test existing agents and refine based on results

Tasks:

  • Test Event Follow-Up Agent with real event data
  • Test LinkedIn Sequence Agent with real prospects
  • Collect feedback on agent outputs
  • Refine templates and selection logic
  • Add more playbooks/sequences

Priority: High (validate quality before building more)

Phase 3: Additional GTM Agents

Goal: Build more specialized GTM agents

Planned Agents:

  1. Prospect Research Agent 🔜

    • Purpose: Research prospects and find personalization angles
    • Input: Prospect name, company, LinkedIn (optional)
    • Output: Research report, personalization angles, qualification signals
    • Knowledge Sources: bdr-tactics.md, qualification-criteria.md
    • Status: Not Started

  2. Outreach Email Agent 🔜

    • Purpose: Generate personalized cold outreach emails
    • Input: Research report, prospect info, personalization angles
    • Output: Personalized email (subject + body), follow-up sequence suggestions
    • Knowledge Sources: message-templates.md, bdr-tactics.md, positioning.md
    • Status: Not Started

  3. Lead Qualification Agent 🔜

    • Purpose: Qualify leads based on research/responses
    • Input: Prospect info, research, interaction history
    • Output: Qualification decision, reasoning, score
    • Knowledge Sources: qualification-criteria.md, bdr-tactics.md
    • Status: Framework exists (ICP Analysis), needs full agent

  4. Response Handling Agent 🔜

    • Purpose: Generate responses to prospect replies/objections
    • Input: Prospect response, context, conversation history
    • Output: Personalized response, objection handling, qualification questions
    • Knowledge Sources: message-templates.md, bdr-tactics.md, objection handling (needs to be created)
    • Status: Not Started

  5. Multi-Channel Follow-Up Agent 🔜

    • Purpose: Coordinate follow-ups across email, LinkedIn, SMS (if opted-in)
    • Input: Prospect info, engagement history, preferred channels
    • Output: Multi-channel follow-up sequence with timing
    • Knowledge Sources: linkedin-playbooks.md, message-templates.md, event-follow-up-playbook.md
    • Status: Not Started

Priority: Medium (after Phase 1 & 2)

Phase 4: Integrations & Automation

Goal: Connect agents to external tools and automate workflows

Tasks:

  • Slack integration (see SLACK_DEPLOYMENT_GUIDE.md)
  • CRM integration (sync leads, handoffs, responses)
  • Email client integration (draft to Gmail/Outlook)
  • Calendar integration (meeting scheduling)
  • Analytics dashboard (track agent performance)

Priority: Low (after agents are validated and refined)

Phase 5: Advanced Features

Goal: Add advanced capabilities based on learnings

Planned Features:

  • A/B testing framework for agent outputs
  • Long-tail follow-up automation (3+ weeks post-event)
  • Predictive qualification (ML-based scoring)
  • Conversation intelligence (analyze call transcripts)
  • Multi-language support
  • Industry-specific agents (home services, SaaS, CPG, etc.)

Priority: Low (future enhancements)

📝 Session Log

Session 1: January 16, 2025

What We Built:

  1. ✅ Event Follow-Up Agent - Complete Alice-style event follow-up system
  2. ✅ LinkedIn Sequence Agent - Multi-sequence playbook system
  3. ✅ ICP Analysis Framework - “Thinking slow, thinking fast” methodology
  4. ✅ Knowledge Base Structure - ICP, qualification, positioning, playbooks
  5. ✅ Slack Deployment Guide - Framework for deploying agents to Slack
  6. ✅ Documentation - Comprehensive tickets, examples, workflows

Key Decisions:

  • Focus: GTM agents only (BDR/outreach), not PM/content/ops
  • Approach: Human-in-the-loop (draft for review, not auto-send)
  • Model: Quality over automation
  • Workflow: Always pull from main before starting new work

Knowledge Captured:

  • ICP Operating System (from Notion) → ideal-customer.md, qualification-criteria.md
  • Vertical testing framework (home services example) → vertical-testing.md
  • LinkedIn playbooks (Job Posting, Mutual Intro) → linkedin-playbooks.md
  • Event follow-up playbook structure → event-follow-up-playbook.md

Knowledge Gaps Identified:

  • BDR tactics not yet documented (structure exists)
  • Message templates not yet populated (structure exists)
  • Positioning not yet documented (structure exists)
  • Client pain points need to be added

Next Session Focus:

  • Fill knowledge gaps (BDR tactics, templates, positioning)
  • Test Event Follow-Up Agent with real event data
  • Build Prospect Research Agent
  • Build Outreach Email Agent

📊 Progress Tracking

Agents Status

AgentStatusTestingIntegration
Event Follow-Up Agent✅ Complete🔄 Ready for Testing❌ Not Started
LinkedIn Sequence Agent✅ Complete❌ Not Started❌ Not Started
ICP Analysis Agent✅ Framework✅ Example Exists❌ Not Started
Prospect Research Agent❌ Not Started--
Outreach Email Agent❌ Not Started--
Lead Qualification Agent🔄 Framework Only--
Response Handling Agent❌ Not Started--
Multi-Channel Follow-Up Agent❌ Not Started--

Knowledge Base Status

Knowledge FileStatusNeeds Content
ICP Profiles✅ Complete-
Qualification Criteria✅ Complete-
Vertical Testing✅ Complete-
Event Follow-Up Playbook✅ Complete-
LinkedIn Playbooks✅ Complete-
BDR Tactics🔄 Structure Only✅ Yes
Message Templates🔄 Structure Only✅ Yes
Positioning🔄 Structure Only✅ Yes

🎯 Success Metrics

Current Goals:

  1. Quality: Agent outputs pass human review (no generic content, proper personalization)
  2. Efficiency: Reduce time to draft outreach by 50%
  3. Consistency: Maintain Brainforge’s unique voice and approach
  4. Coverage: Support all major GTM workflows (event follow-up, cold outreach, qualification)

Future Goals:

  1. Response Rates: Match or exceed human-generated outreach
  2. Qualification Accuracy: 80%+ accuracy in lead qualification
  3. Pipeline Generation: Generate X qualified leads per month
  4. Time Savings: Free up 10+ hours/week for team members

📚 Related Documentation

  • Agent Scaling Guide: ../AGENT_SCALING_GUIDE.md
  • GTM Agent Sprint: ../GTM_AGENT_SPRINT.md
  • Quick Start: ../QUICK_START.md
  • Slack Deployment Guide: gtm/agents/SLACK_DEPLOYMENT_GUIDE.md
  • Notion to Vault Mapping: gtm/agents/NOTION_TO_VAULT_MAPPING.md

🤝 Contributing

How to Contribute:

  1. Review roadmap and identify gaps
  2. Fill in knowledge base (BDR tactics, templates, positioning)
  3. Test existing agents and provide feedback
  4. Build new agents following existing patterns
  5. Update this roadmap with progress

Questions?

  • Check agent documentation in gtm/agents/.claude/agents/
  • Review examples in gtm/agents/examples/
  • Read knowledge base in gtm/agents/memory/

Last Updated: 2025-01-16
Next Review: After Phase 1 (Knowledge Capture) completion

Graph View

Backlinks