Bricks And Bytes What Construction AI Is Missing: Connecting the Dots With Knowledge Graphs
8 snips
Oct 22, 2026 
Bogdan Arsintescu, an original Google Knowledge Graph team member who later led graph systems at LinkedIn and Microsoft, joins Zaid Kahn, co-founder of Neuron Factory, to unpack construction AI’s missing layer. They explore knowledge graphs, ontologies, document conflicts, addenda, and why generic chatbots lose crucial context. The conversation also examines vertical AI, preconstruction risk, and preserving estimators’ hard-won project knowledge.
AI Snips
Chapters
Transcript
Episode notes
Ontologies Define Construction’s Rulebook
- An ontology defines what objects and symbols mean, while the graph connects those defined entities across project documents.
- A door can link its specification, drawing schedule, and RFI so changes become traceable across sources.
Graphs Expose Conflicts Across Drawings
- A knowledge graph maps the relationships among a project’s door schedule, specifications, drawings, and regulations.
- It can detect that a door’s 45-minute rating violates a city requirement for two-hour external fire doors.
The Addendum Hidden From The Estimator
- An estimator prices doors from a specification and unchanged schedule while a separate addendum quietly changes their ratings.
- The graph links all three sources, surfaces conflicting values, and identifies which document supersedes another.