Treadstone Associates
Ask an Expert · 4 min read

What if the crew built from an old drawing revision?

This isn't a statute question -- it runs through the contract's change-management channel. But before anyone argues fault, you need to prove which set was actually current that day.

Treadstone Associates · Updated 2026

Short answer

Who pays comes down to who let the stale set stay in circulation -- but before that argument starts, the more useful question is how fast you can prove which revision was actually current on the day the work happened.

Prove the timeline before you argue the fault

Bluebeam's Compare Documents tool exists for exactly this. It “highlights differences between two PDFs using cloud markups”, producing a new file with differences between the original and revised drawings highlighted in orange, without altering either source document. Run the superseded set against the current one and you have a dated, visual record of exactly what changed -- a much stronger starting point than a dispute over who forgot to update a drawing rack. A companion feature, Overlay Pages, layers multiple versions with a distinct colour per file, useful when more than two revisions are in play.

Track revisions as part of the daily record, not a separate system

Procore's Daily Log includes a dedicated Plan Revisions section that “allows you to add tracking information about project plans by revision, title, and category”, so a superseded set showing up on site gets logged the day it's caught, rather than reconstructed from memory weeks later once it's already a dispute.

Where the liability question actually lands

No statute assigns fault for this automatically -- it runs through the contract's change-management and RFI channels, which is why the response time on a related RFI matters (see how long should an RFI take to answer): if the crew asked and the answer came back late, or never came, that shifts the story. If the current set was issued and available and the crew built from an old one anyway, that's a different conversation entirely. Either way, document control -- who issued what, when, and who acknowledged receiving it -- is what decides which conversation you're actually having.

The cheap fix is upstream of the dispute

Most old-revision incidents trace back to the same root cause: a printed or downloaded set that never got recalled when a newer one issued. A sheet index with a revision column, checked against the current issue log before a crew starts a task, catches most of this before it becomes rework -- and it's a lighter lift than reconciling a built condition against a superseded drawing after the fact. Treat every reissue the same way a change order gets treated: dated, distributed, and acknowledged, not just uploaded somewhere and assumed to have been seen.

Catch a stale drawing before it becomes a change order.

A 30-minute call is enough to tell you whether AI pays for itself here.