Neuron Factory

Neuron Factory starts from a data structure. Before any agent writes a scope or flags a conflict, the platform builds a graph of how a project's documents relate: which spec section governs which drawing note, which revision supersedes which, which bid covers which line item. The team behind it built graph systems at LinkedIn, Google, and Microsoft, and the pitch is that preconstruction errors live in the relationships between documents, so the relationships need to be modeled explicitly rather than recovered by search each time.

In practice the product shows up in two places a preconstruction manager recognizes. Project Graph handles bid packaging: it scans the full set, pulls the sheets and spec sections for each trade, keeps a running view of scope nobody has been assigned yet, writes a Word scope document in the contractor's format, and levels the quotes that come back. Conflict Graph is the constructability side, comparing drawings and specifications across disciplines and ranking what it finds by severity so the team works the expensive problems first. Both are aimed at general contractors in the weeks before bid, and the company says it tests on live projects with GCs in North America, Europe, and the Middle East.

Buyers should test how the graph handles addenda and reissued sheets, whether conflict severity rankings match what their estimators would flag, and how easily the scope output drops into their bid management system. Nomic works on overlapping problems, including cross-discipline drawing review and scope gaps in bid sets ranked by cost risk, while bid leveling and scope package authoring are Neuron Factory's focus.