Navigate Knowledge

Building Knowledge Graphs where every entity is a handle back to the evidence that created it.

WORKSHOP AGENDA

Building Knowledge Maps: Evidence-First Knowledge Graphs for Unstructured Data

May 5, 2026
3:30 – 5:00 PM

Hands-on Workshop (90 Min)

JIT-Resolution: From Symbolic Graphs to Contextual Intelligence for Trustworthy AI Reasoning

May 7, 2026
11:05 – 11:30 AM

Conference Talk - Technology (25 min)

image of a graph filling the Kineviz' GraphXR interface

THE APPROACH

Entities as handles. Not facts.

Most knowledge graphs collapse uncertainty into confidence. Knowledge Maps keep the thread back to evidence - so AI reasons from source, not schema.

- Evidence-first. Every node links to the exact span in the source document.

- Just-In-Time Resolution. Entity resolution is deferred to query time. Ambiguity is represented,

not erased.

- SQL is ETL. Extraction and graph construction run inside BigQuery — no pipeline rebuilds, no

data movement.

A mind map showing John, who was a Model Risk Governance Lead at JPMorgan Chase, working at JPM and reporting to the Chief Risk Officer from 2018 to 2022 in New York, with an observation box describing his role.

Built on GraphXR

Learn more about knowledge graphs →