Owen School of Management · AI-Accelerated Entrepreneurship · Spring 2026

The Trinity Graph
Cohort Architecture

How the entire class connects into a collective intelligence network — and what serendipity looks like when it works.

SOCIAL · KNOWLEDGE · GENERATIVE · SERENDIPITY ENGINE
The Big Picture

What the Trinity Graph Actually Is

The Trinity Graph is not a tool. It is a biological operating system layered over the entire cohort. Think of every student as a node. Every conversation, every project, every connection is an edge. The graph is the living map of the class's collective intelligence — and the serendipity engine is the algorithm that traverses it to surface connections nobody would have found on their own.

There are three layers. They are not separate products. They are the same system at three different levels of abstraction.

◆ KNOWLEDGE

WHAT — The Information Graph

Every domain, skill, market, dataset, and insight each student carries. Triples: Entity → Relationship → Entity. Nashville R1 → requires → 5ft setback. Healthcare → regulated by → FDA.

✦ GENERATIVE

WHAT IF — The Synthesis Engine

The layer that traverses both graphs to produce something new. A feasibility score. A warm intro path. A market hypothesis. The output no individual could have generated alone.

⟁ SERENDIPITY

THE UNEXPECTED — Cross-Graph Bridges

The emergent layer. When the algorithm finds a connection between two nodes that nobody programmed, nobody anticipated, and nobody would have found through normal networking. This is the prize.

Graph Architecture

What Gets Stored as a Node

Every student's individual knowledge graph feeds into the collective. Here are the node types that exist across the cohort map.

● SOCIAL NODE

Student / Person

Name, background, archetype (Connector, Builder, etc.), Kirk Step, spirit animal. The base identity node.

● SOCIAL NODE

Warm Network Contact

First-degree connections each student contributes. GC uncle, title insurance contact, city planner friend. These are the edges that bridge structural holes.

◆ KNOWLEDGE NODE

Skill / Domain

What each student actually knows. Supply chain, regulatory compliance, data science, sales. Weighted by depth (novice → expert).

◆ KNOWLEDGE NODE

Market / Industry

The domains each venture operates in. PropTech, Healthcare, FinTech, Art. Connects ventures to relevant student expertise automatically.

✦ GENERATIVE NODE

Venture / Project

Each pod's startup. BackyardOne, Artiquity, Enginuity, Aura Health, Ubiquity. The unit that the whole graph is trying to serve.

✦ GENERATIVE NODE

Insight / Discovery

A customer discovery finding, a validated hypothesis, a market signal. These are the artifacts that emerge from pod interviews and feed back into the graph.

⟁ BRIDGE NODE

Structural Hole

A detected gap between two clusters. The system flags these automatically. "BackyardOne needs regulatory expertise. MBA-3 has it but is on the Artiquity pod."

⟁ BRIDGE NODE

Serendipity Event

A surfaced connection that was unexpected. A cross-pod insight. A warm intro nobody knew existed. The graph's highest-value output.

The Core Innovation

How the Serendipity Engine Works

Serendipity in a recommender system means surfacing connections that are unexpected but relevant — not random noise, not obvious matches. The serendipity engine is what separates the Trinity Graph from a CRM or a LinkedIn graph.

⟁ SERENDIPITY ENGINE — 5-STEP TRAVERSAL

1

Map Every Student's Latent Interest Graph

Not just stated preferences. The system infers latent interests from project choices, collaboration patterns, and conversation data. A student who says "fintech" but keeps asking questions about regulatory arbitrage has a latent interest in policy.

2

Detect Structural Holes Across Pods

The algorithm scans for gaps between venture clusters. BackyardOne needs healthcare-adjacent regulatory experience? The graph checks every student across all pods, not just the BackyardOne team.

3

Calculate Serendipity Score

Each potential connection gets a score: how unexpected is it (novelty) × how relevant is it (utility). Score too low = boring and obvious. Score too high = random noise. The sweet spot is high novelty + high relevance.

4

Surface via the Membrane (VanderBot)

The engine doesn't email you a list. It surfaces the connection at the right moment — during a relevant conversation, before a session where cross-pod collaboration would be valuable, or when a venture hits a specific kind of wall.

5

Write the Outcome Back into the Graph

If the connection was acted on — a collaboration happened, a warm intro was made — the graph records the edge with weight. Over time, this teaches the serendipity engine which patterns generate real value vs. false positives.

What Ideal Serendipity Looks Like in Practice

CROSS-POD SKILL BRIDGE

Artiquity needs blockchain provenance expertise. The graph detects that a BackyardOne MBA has prior experience with property title chains on distributed ledgers. VanderBot surfaces the connection before Demo Day. Neither pod knew this existed.

WARM INTRO SHORTCUT

BackyardOne needs access to Nashville zoning officials. The graph finds that an Ubiquity pod member's family member works in Metro Planning. Three degrees of separation → one warm intro. CAC = $0.

LATENT MARKET SIGNAL

Three different pods, doing customer discovery in completely different markets, all independently hear the same friction point: "the permit process is broken." The graph aggregates these signals and flags it as a cross-market structural hole worth investigating.

COMPLEMENTARY ARCHETYPE MATCH

Enginuity's pod is all Builders. They are technically excellent but stuck on go-to-market. The graph detects that Adele (The Connector) has solved a structurally similar GTM problem. VanderBot suggests a 20-minute cross-pod session.

Your Position in the Graph

The BackyardOne Pod Architecture

Your pod is a micro-instance of the Trinity Graph. Adele is the Brain (topology). The 4 MBAs are the Membrane (sensory organ). The graph is the infrastructure that makes their outbound energy compound rather than evaporate.

🧠

Adele

Systems Architect

BRAIN — Topology
👁

MBA 1

Customer Discovery

MEMBRANE — Sensing
👁

MBA 2

Network Seeding

MEMBRANE — Sensing
👁

MBA 3

Market Validation

MEMBRANE — Sensing
👁

MBA 4

Network Seeding

MEMBRANE — Sensing

Each MBA is a node contributing their warm network edges to the graph. When they feed contacts back to Adele, those nodes get wired in. The graph grows. The serendipity engine has more to work with. The pod's collective reach expands exponentially.

Scale Architecture

How the Graph Compounds Over Time

The Trinity Graph is designed to get more powerful as more nodes and edges enter the system. Here is the scaling arc across the course:

WK 1-2

Individual Nodes — WHO

Each student enters the graph as a node. Basic attributes: archetype, background, Kirk Step. VanderBot begins learning each student's individual knowledge graph through conversation.

WK 3-4

Knowledge Triples — WHAT

Each pod builds their 100-triple Knowledge Graph. These triples get ingested into the collective system. Now the graph knows not just WHO each student is but WHAT they know.

WK 5-7

Generative Traversal — WHAT IF

The serendipity engine begins running cross-graph traversal. The system starts detecting structural holes across pods. VanderBot surfaces the first unexpected cross-pod connections. The graph starts generating value it was not explicitly programmed to generate.

SESSION 7

Midterm — The First Full Map

The Laptop Drop is the first moment the entire class can see the collective graph. All five Inkwell ventures mapped. All warm networks visible. All structural holes identified. The serendipity engine goes live across the full cohort.

SESSION 14

Demo Day — The Organism Is Alive

By Demo Day, the graph has ingested weeks of customer discovery, cross-pod collaboration, and real market signals. Each venture's pitch is no longer just a slide deck — it is a node in a living network that can demonstrate compounding value in real time.

The Vision

What the Ideal State Looks Like

When the Trinity Graph is fully operational across the cohort, this is what becomes possible:

NO COLD OUTREACH

Every customer discovery interview, every investor conversation, every advisor relationship starts with a warm intro traced through the cohort's collective warm network graph. Zero cold outreach for any pod.

ZERO KNOWLEDGE LOSS

Every customer interview, market signal, and validated hypothesis gets written back into the graph as a node. Nothing lives in a Google Doc or a forgotten Slack thread. The cohort's collective intelligence is permanently accessible and queryable.

EMERGENT VENTURES

The cross-graph traversal surfaces a market opportunity that nobody was explicitly looking for — a structural hole visible only when two pods' customer discovery data is analyzed together. A new venture is born from the graph's output, not from a founder's isolated idea.

SELF-REINFORCING NETWORK

Metcalfe's Law: network value = N². Each new node (student, contact, insight) added to the graph increases the total value for every other node. By Demo Day, the graph is worth more than the sum of its parts. That is the organism.

SOURCES & CITATIONS