The Moat · The Engine

The Context Graph: decades of real business knowledge, encoded.

Most tools tell you what's complex. We connect the why behind it — and the how to solve it.

What's in it

Four assets. One graph.

Decades
Real business knowledge
Cases, patterns, root causes, what worked, what didn't — encoded.
Seeded
From real engagements
User contribution opens at public beta. Seed count and provenance ship with MVP.
386
Anchored intersections
The 6 × 4 × ~16 backbone is the taxonomy layer.
Traceable
Every recommendation
Each output cites the precedents and the pathway.
The Graph

Click a node. Trace its edges. Read its precedents.

Five node types. Six edge types. Every output is a traversal through this graph, anchored on the 386 backbone.

The Schema

Five node types. Six edge types.

For technical readers — and for engineers integrating the graph.

Node types
Pattern Cause Intervention Outcome Precedent
Every node is tagged with Principle × Dimension × Sub-Theme. The 386 backbone is the index — not the body.
Edge types
causes mitigated_by predicts instance_of observed_in similar_to
Edge weights are continuously updated by outcome telemetry — strong outcomes pull future recommendations toward proven interventions.
How it grows

Ingest. Retrieve. Evolve.

Closed-loop learning, built in from day one.

01

Ingest

  1. Consulting case or user contribution arrives
  2. LLM extracts entities + relations
  3. Reconciliation: dedupe, merge, link
  4. Populate graph + embed for retrieval
02

Retrieve

  1. Question → embed → anchor on 386
  2. Graph traversal (Cypher) + vector hybrid
  3. Re-rank by precedent strength
  4. Synthesize → cite nodes + edges
03

Evolve

  1. Every assessment adds nodes and edges
  2. 3 / 6 / 12-month outcome telemetry logged
  3. Edge weights re-scored continuously
  4. Contributions anonymized and curated before commit
Defensibility

Why it can't be cloned in less than a decade.

The graph is not a dataset you license. It is the compounding artifact of decades of paid consulting practice, encoded one node at a time — and built to compound further with every user assessment once contribution opens at public beta. A competitor cannot replay the work, cannot rebuild the institutional pattern memory in a sprint, and cannot reproduce the outcome telemetry that re-weights every edge.

The engines on top are the products. The graph underneath is the asset.

Two doors into the Forge

Pick the door that fits where you are.

Free general membership, or the paid contributor tier. Two ways into the community that practices Simplexitude — nothing else on the site requires either.

Door 1 · Free · by application

Join the Forge.

General membership is free. Events, peer circles, workshops, calendar of gatherings. You don't pay; you don't have to buy the product.

Join the Forge — free →
Door 2 · Paid contributor · 40 of 50

Founding Council.

Shape the engines, not just use them. 2–4 hours / month + real cases. In return: first MVP access, $2,500 credits, $5–10k success share, named seat.

Apply to the Council →