Distinction Infrastructure — The Jarvis Memory Layer

Chapter 21 Companion

Distinction
Infrastructure

The Jarvis Memory Layer — not a contact list, but a structured knowledge base of every meaningful relationship and interaction. Query "Tell me everything you know about Sarah Chen" and receive a synthesized briefing of every interaction, topic, and commitment.

CONTACTS Table

The foundation — who matters to you, how you met, and what makes them memorable. Includes AI-generated mnemonic descriptions and physical characteristics for face recognition assistance.

Schema

- Basic info (name, email, phone, company)
- Relationship context (how we met, shared interests)
- Mnemonic data (AI-generated memorable descriptions)
- Physical characteristics (for face recognition assistance)
- Last interaction, next touch due

INTERACTIONS Table

Every meaningful exchange — meetings, calls, emails — with summaries, key points, and follow-up commitments. The raw material that Jarvis synthesizes into relationship intelligence.

Schema

- Every meaningful exchange (meetings, calls, emails)
- Summaries and key points
- Location data (where we met)
- Follow-up commitments
- Topics discussed

VECTOR_CHUNKS Table

Semantic embeddings of all notes and content — enabling search by meaning, not just keywords. Includes spatial coordinates for location-tagged memories and hierarchical relationships between documents.

Schema

- Semantic embeddings of all my notes and content
- Spatial coordinates (location-tagged memories)
- Hierarchical relationships (parent documents, chunks)
- Cross-references to contacts and interactions

How the three tables work together

CONTACTS stores who matters. INTERACTIONS stores what happened. VECTOR_CHUNKS enables semantic search across everything — including location-aware queries that surface relevant memories based on where you are, not just what you're thinking about.

From AI in Business Strategy by Sean M. Bair.
Get the book — or explore more companion resources.

Copied!