Farouk — Discovery Intelligence Package

Prospect Profile — Farouk

Overview

Farouk is a marketing agency owner specializing in media buying and SEO, primarily serving the pet industry. He operates multiple businesses and has been building AI workflows independently using consumer and prosumer tools. Alexander assessed him as being "where I was about four and a half months ago" in his AI journey — technically curious, actively experimenting, but not yet operating at the infrastructure level MERIDIAN provides.

Business Portfolio

Primary: Marketing Agency

•Focus: Media buying and SEO
•Vertical: Pet industry
•Key IP Source: Mario Castelli's copywriting frameworks and course materials ($500/mo subscription)
•Approach: Data-driven, uses historical performance data to inform campaigns

Secondary Businesses

•2-3 additional businesses "on the backburner"
•Neglected due to bandwidth — running the agency consumes most of his time
•Specific verticals not disclosed in discovery call
•These represent potential MERIDIAN use cases if trust develops

Proprietary IP Assets

Farouk is protective of his intellectual property — described as having "300 million plus resources" across:

Asset Type	Platform	Notes
Campaign data	Google Sheets	Historical performance data, metrics, benchmarks
Frameworks	Google Docs	Copywriting frameworks, SOPs, process documentation
Visual assets	Screenshots	Ad creatives, performance dashboards, competitive analysis
Course materials	Downloads	Mario Castelli CRO expert repos, copywriting courses
Data points	Various	Industry benchmarks, pet industry market data

Trust Gate: Farouk explicitly stated he's not ready to share via Google Drive yet. The "speed of trust" principle applies — relationship must develop before IP sharing.

Personality & Working Style

•Analytical: Rates things, measures performance, thinks in frameworks
•Builder: Already designing multi-agent architectures on his own
•Protective: Guards his IP carefully, needs to see value before sharing
•Time-constrained: Good ideas but limited execution bandwidth
•Self-directed learner: Downloads repos, watches courses, experiments with tools
•Pop culture references: Named his AI assistant "Donna" after the Suits character

Relationship Context

•Connected through Michael Bitler's network
•First discovery call — no prior relationship with Alexander
•Michael serves as the trust bridge between parties
•Next interaction needs to demonstrate competence through pre-research on his businesses

Current Tech Stack & AI Readiness

Current Tools

Tool	Usage	Proficiency
ChatGPT	Primary AI assistant	Daily user
Claude Desktop	Secondary AI, exploring	Experimenting
N8N	Workflow automation	Building flows
Google Sheets/Docs	Data storage, documentation	Power user
GitHub	Downloading repos (CRO expert, Castelli materials)	Consumer level
Open Router	Planned — model routing	Researching

Planned Architecture (Pre-MERIDIAN)

Farouk was independently designing an AI agency architecture that mirrors MERIDIAN's approach:

The Mac Mini Cluster Plan

•Concept: Individual Mac Minis, each assigned a specialized role
•Roles planned:
•Graphic Designer agent
•Video Editor agent
•Social Media Manager agent
•Copywriter agent (powered by Castelli frameworks)
•Routing: Open Router to select best model per task
•Chief of Staff: AI named "Donna" to coordinate all agents
•Spent 8-10 hours building the prompt
•Modeled after Donna Paulsen from Suits (anticipates needs, manages workflow)

Architecture Assessment

Aspect	Farouk's Plan	MERIDIAN Equivalent
Compute nodes	Mac Minis (isolated)	Mesh network (redundant, coordinated)
Agent coordination	Single "Chief of Staff" prompt	Orchestrator with coherence gate
Memory	Conversation history only	Persistent graph (Neo4j MIG, 17K+ nodes)
Model routing	Open Router (manual config)	Cost Router with budget-aware tier selection
Failure handling	Single point of failure per box	Circuit breakers, DLQ, mesh failover
IP ingestion	Manual copy-paste into prompts	Automated research extraction pipeline

Key insight: Farouk is solving the same problems MERIDIAN already solves. His architecture would work for a single operator but lacks the persistence, redundancy, and coherence layers that make MERIDIAN enterprise-grade.

AI Maturity Assessment

Level: Intermediate (3 out of 5)

•Understands the concept of specialized agents
•Knows about model routing and API patterns
•Has downloaded and explored code repos
•Building N8N workflows for automation
•Missing: persistent memory, mesh coordination, deployment infrastructure, coherence/quality gates

Gap Analysis:

1.No persistent memory across sessions — loses context between conversations
2.No quality control layer — no equivalent to coherence gate or noise gate
3.Single-box architecture — no redundancy if a Mac Mini fails
4.No automated ingestion pipeline — manual copy-paste for IP
5.No deployment infrastructure — no CI/CD, no blue-green, no rollback

Technology Mentioned in Call

Sesame AI (sesame.com)

•Voice AI platform in beta
•Extremely human-like conversational voice
•Free 5-minute calls during beta
•Potential evaluation target for Agent Vox voice stack comparison

Mario Castelli / CRO Expert

•Copywriting and conversion rate optimization frameworks
•$500/month subscription for course materials
•Farouk has downloaded repos and course content
•Already partially ingested into his Claude conversations

Discovery Call Debrief

Date: March 29, 2026 Duration: ~45 minutes Participants: Farouk (prospect), Michael Bitler (business development), Alexander (MERIDIAN) Format: Video call with screen sharing

Call Grade

Grader	Score	Rationale
Michael Bitler	C+ to B-	Lacked pre-research; didn't have Farouk's business details ready
Alexander	C	Same — the call should have been a prepared second meeting, not a cold discovery

Root cause of low grade: No pre-research was conducted on Farouk's businesses before the call. The team went in without knowing his websites, SEO footprint, or competitive landscape. This meant the conversation was generic exploration rather than demonstrating MERIDIAN's value through specific, targeted insights about *his* business.

What Was Demonstrated

MERIDIAN Capabilities Shown

1.Boardroom meetings — Multi-daemon voice conversations with distinct personalities
2.Telegram bot — @OracleDelphiBot real-time interaction
3.Recording pipeline — Live transcription from call recordings
4.Memory mesh — Persistent knowledge across sessions (MIG)
5.Coherence gate — Sensei quality control on daemon outputs
6.Noise gates — Filtering low-signal information
7.Email orchestration — Unified Gmail catchall with automated routing
8.Redundancy architecture — Mesh vs. single-point-of-failure comparison

What Landed

•The concept of specialized daemons resonated — maps directly to his Mac Mini plan
•Persistent memory across sessions — he recognized this gap in his own setup
•The redundancy argument — "what happens when your one Mac Mini dies?"
•The speed of execution — MERIDIAN is already built and running

What Didn't Land

•Hard to assess impact without seeing his actual business data
•Generic demo without customization to his specific problems
•Trust barrier prevented deeper engagement on IP ingestion

Key Moments

"Where I Was 4.5 Months Ago"

Alexander recognized Farouk's trajectory as matching his own earlier journey. The implication: Farouk could build this himself given enough time, but MERIDIAN represents a 4+ month head start with production-grade infrastructure.

"Donna" — The Chief of Staff

Farouk spent 8-10 hours crafting a single AI assistant prompt modeled after Donna Paulsen from Suits. This shows:

•He's already thinking about AI personality and role design
•He's investing significant time in prompt engineering
•His approach is single-agent (one mega-prompt) vs. MERIDIAN's multi-daemon mesh

IP Protection Stance

When the topic of sharing Google Drive came up, Farouk pushed back. He has "speed of trust" instincts — won't share proprietary data until the relationship proves safe. This is a *healthy* signal for a potential JV partner (protects IP) but means the sales cycle requires patience.

The Mac Mini Architecture

Farouk was planning to buy Mac Minis and assign each one a role. This is architecturally similar to MERIDIAN's daemon model but without:

•Shared memory
•Coordination layer
•Quality gates
•Deployment infrastructure
•Failure recovery

The overlap is the strongest sales angle: "You're building what we've already built."

Competitive Positioning

Against Building It Himself

Factor	DIY (Farouk's plan)	MERIDIAN
Time to operational	3-6 months minimum	Already running
Memory persistence	None (session-based)	17,900+ nodes in MIG
Agent coordination	Manual / single prompt	Automated orchestrator
Quality control	None	Coherence + noise gates
Deployment	Manual per machine	Blue-green, zero-downtime
Cost	Hardware + time + iteration	Subscription / JV revenue split

Against Other AI Platforms

Farouk hasn't mentioned evaluating other enterprise AI platforms. His frame of reference is consumer tools (ChatGPT, Claude Desktop) and DIY (N8N, Mac Minis). MERIDIAN occupies a different tier.

Partnership & Corporate Structure

Context

The discovery call included significant discussion about how domain experts like Farouk could participate in the MERIDIAN ecosystem without diluting core founder equity. Michael Bitler outlined a JV subsidiary model that balances participation incentives with corporate protection.

Proposed Corporate Architecture

`` CoVault (Delaware C-Corp) Parent Entity — Core IP, Platform Founders: ~20% each x 3 = 60% │ ├── Puerto Rico LLC (Act 60) │ Tax-optimized operating entity │ ├── JV Sub: Marketing Domain (Farouk) │ Revenue split on marketing vertical │ Farouk contributes: IP, domain expertise, client relationships │ MERIDIAN contributes: platform, infrastructure, AI capabilities │ ├── JV Sub: Medical/HIPAA (Omar) │ Revenue split on healthcare vertical │ Omar contributes: compliance expertise, medical network │ ├── JV Sub: Manufacturing (Alan) │ Revenue split on manufacturing vertical │ Alan contributes: green data center, university connections │ └── Future JV Subs (per vertical) Scalable model for additional domain experts ``

Key Principles

1. Core Equity Stays Tight

•3 co-founders hold ~60% combined
•Domain experts do NOT get founder-level equity
•Participation is through revenue splits, not ownership dilution

2. JV Subsidiary Model

•Each domain expert gets a subsidiary entity
•Revenue split (discussed range: 5-10%) on their vertical
•They contribute: IP, domain knowledge, client relationships, industry credibility
•MERIDIAN contributes: platform, infrastructure, AI agents, ongoing development

3. Corporate Isolation

•Each subsidiary is legally separable
•If one vertical fails, it doesn't damage the parent or other subs
•Clean shutdown path if a partnership doesn't work out

4. IP Protection (Both Directions)

•Farouk's IP stays attributed to him
•MERIDIAN's platform IP stays with CoVault
•Clear licensing boundaries in JV agreement
•Work product from the AI agents — ownership defined per agreement

Farouk-Specific Considerations

What Farouk Brings

Asset	Value	Notes
Marketing agency operations	Domain expertise	Media buying, SEO, campaign management
Pet industry network	Client pipeline	Established relationships in vertical
Mario Castelli frameworks	Copywriting IP	Course materials, CRO frameworks
Historical campaign data	Training data	300M+ resources in Sheets/Docs
Technical capability	Self-sufficiency	Can co-develop, not just consume

What Farouk Needs

Need	MERIDIAN Solution
More bandwidth	AI agents handling routine tasks
Persistent context	MIG memory across all conversations
Coordinated agents	Mesh orchestrator vs. single Donna prompt
Redundancy	Multi-node architecture vs. single Mac Mini
Scale	Enterprise infrastructure already built

Trust Timeline

Farouk is not ready for a JV conversation yet. Current position:

1.Now — Discovery phase, building familiarity
2.Next call — Demonstrate competence with pre-researched insights on his businesses
3.After that — Specific value demonstration (small proof-of-concept with his data)
4.Then — JV structure conversation with Michael present
5.Close — Formal agreement, IP ingestion, subsidiary setup

Alan

•Asset: Green data center, university connections (professors + students for maintenance)
•Vertical: Manufacturing (related to Bezos $100B automation announcement)
•Contribution: Infrastructure + Sensei upgrades to MERIDIAN
•Model: Same JV subsidiary structure

Omar

•Asset: Medical/HIPAA compliance expertise
•Vertical: Healthcare
•Model: Same JV subsidiary structure
•Key differentiator: Compliance knowledge is the moat in healthcare AI

Tony Moore

•Background: Ex-Microsoft, based in North Carolina
•Potential role: Sales/contact network contributor
•Status: Michael to introduce
•Model: IC agreement (sales partner), not JV

Legal Considerations Raised

Contract Feedback (from Legal GPT Review)

The Agent Vox contract iterations surfaced these concerns applicable to all JV agreements:

•Tier 3 licensing may be too generous — allows sublicensing, derivative works, assignment to acquirers
•Needs tighter audit, attribution, and underpayment remedies
•IP assignment vs. work-for-hire distinction needs clarity per engagement
•Ely and Alexander to review iterations before any new agreements go out

Action Items & Next Steps

Priority Matrix

Critical Path (Must Complete Before Next Call)

#	Action	Owner	Status	Notes
1	Get Farouk's business websites/URLs	Alexander via Michael	Pending	Cannot pre-research without this
2	Run full prospect intelligence on Farouk's businesses	MERIDIAN (Atlas/Ronin)	Blocked by #1	Company profile, SEO audit, competitive landscape, digital presence — same depth as Rank On Amazon package
3	Schedule second discovery call	Michael	Pending	Should feel like "the call the first one should have been"

High Priority (This Week)

#	Action	Owner	Status	Notes
4	Follow up with Denver Griffin for data/IP	Michael	Pending	Has not provided data despite prior discussions
5	Get back to Brian Cook	Michael	Pending	Mentioned as open item
6	Set up IC agreement templates for sales partners	Michael / Legal	Pending	Needed for Tony Moore and other network contributors
7	Review Agent Vox contract iterations with Ely	Alexander + Ely	Pending	IP assignment vs. work-for-hire, Tier 3 licensing concerns

Medium Priority (Next 2 Weeks)

#	Action	Owner	Status	Notes
8	Evaluate Sesame AI voice platform	Delphi / Forge	Pending	Beta is free — compare against Agent Vox voice stack
9	Explore Tony Moore intro (ex-Microsoft NC)	Michael	Pending	Sales/contact network potential
10	Prepare JV subsidiary term sheet template	Michael / Legal	Pending	Reusable across Farouk, Omar, Alan partnerships

Low Priority (When Trust Develops)

#	Action	Owner	Status	Notes
11	Small proof-of-concept with Farouk's data	MERIDIAN	Future	Demonstrate value with a subset of his campaign data
12	Mario Castelli framework ingestion	Atlas	Future	If Farouk shares copywriting materials
13	Formal JV subsidiary setup for marketing vertical	Michael / Legal	Future	After trust-building and POC success

Next Call Preparation Checklist

Before the next call with Farouk, ensure:

•[ ] Farouk's primary business website(s) identified
•[ ] Company profile page generated (like Rank On Amazon 01)
•[ ] SEO audit / digital presence analysis completed
•[ ] Competitive landscape mapped for his vertical
•[ ] 3-5 specific insights about his business prepared (not generic)
•[ ] Demo customized to marketing agency use cases
•[ ] "Donna replacement" angle prepared — show how MERIDIAN replaces his Chief of Staff concept with something production-grade
•[ ] Quick wins identified — what could MERIDIAN do for him in week 1?

Meeting Cadence

Meeting	Frequency	Participants	Purpose
Bitler sync	Weekly	Michael + Alexander	Pipeline review, deal strategy
Farouk follow-up	TBD	All three	Second discovery with pre-research
Deal review	As needed	Michael + Alexander + Ely	Contract terms, IP structure

Pipeline Position

``` [DISCOVERY] -----> [PRE-RESEARCH] -----> [SECOND CALL] -----> [POC] -----> [JV STRUCTURE] ^

WE ARE HERE ```

Next milestone: Complete pre-research on Farouk's businesses. This single action transforms the next call from another generic discovery into a competence demonstration.