A reviewer's playbook for verifying that record drawings actually match the building
As-built drawings are supposed to document what was actually constructed, but they only deliver value if someone validates them before accepting closeout. Reviewing as-builts is a separate discipline from creating them: the contractor produces red-lined sets, the architect issues record drawings, and the owner or owner's rep verifies that those documents match physical reality and will be useful for operations, renovations, and future sales.
A thorough review catches missed scope changes, ambiguous markups, and stale drawings before they become a problem 5 years from now when a contractor cuts into a wall expecting empty cavity and hits live conduit.
A complete as-built set should include, at minimum, the following deliverables across every discipline:
Don't trust the paperwork. A defensible review pulls a sample of high-risk items and verifies them against the building.
Focus on items that are buried, expensive to expose, or safety-critical: underground utilities, in-slab conduit, fire dampers, structural connections, and concealed valves.
Mechanical and electrical rooms reveal the largest gap between design and reality. Verify equipment tags, model numbers, and routing matches the record set.
Lift ceiling tiles in a half-dozen representative rooms and compare what you see to the RCP and MEP record drawings.
Pull as-built site utilities and verify against locating marks or a probe survey in at least two critical areas.
Diff the record drawings against the original construction documents using an automated PDF comparison tool. Areas with zero diff in disciplines that you know had RFIs and change orders are a red flag, they almost always indicate missing as-built markups.
Manual page-by-page review of a 600-sheet set is impractical. Useful tools for reviewers include PDF comparison utilities, AI-powered drawing analyzers, BIM model checkers, and field-marking apps that tie photos to drawing locations. For complex projects, a combination of automated difference detection and targeted field walks produces the best results in the least time.
Related references for closeout review and as-built documentation.