Built environment

Our architects, structural engineers, and general contractors bring experience across design and construction, including at Gensler and SOM.
Groundbook AI is built by AI engineers, structural engineers, architects, and general contractors who understand both the complexity of a coordinated drawing set and the rigor required of production software.
Our team collectively brings decades of experience across architecture, structural engineering, general contracting, software, and artificial intelligence.

Our architects, structural engineers, and general contractors bring experience across design and construction, including at Gensler and SOM.
Our engineers bring technology experience from companies including Google and Airbnb.
Our founder is a Y Combinator alum.
Team academic backgrounds include Carnegie Mellon University, University of Michigan, and University of Washington.
Building documents are interconnected. A useful review has to follow relationships across drawings, schedules, specifications, checklists, and code sources—and show the evidence behind every possible issue.
Groundbook brings construction context and AI engineering together in one review workflow, while keeping final judgment with the professionals responsible for the work.
Questions, answered
A team combining AI engineering with practising AEC expertise: the people who build the review system work alongside structural engineers and architects who have run document reviews on real projects. Groundbook AI is a product of Matter0, Inc., based in San Francisco.
No, and it does not act as one. It is software that reviews documents and reports possible issues with the evidence attached. Design, engineering, code, and construction decisions stay with the licensed professionals responsible for the project and with the authority having jurisdiction.
Each finding has to point at specific text or geometry on specific sheets. A second independent pass tries to disprove a finding before it is published, and reviewers' accept and reject decisions feed back into later checks. Findings are framed as questions to verify rather than verdicts.
The sample report walks through nine findings from two published, anonymized project reviews, each shown with its sources and the verification a reviewer would do.