Register intelligence.
Late deliverables against IFC dates, documents stuck in review, revisions issued but unacknowledged by their holders — surfaced continuously, not discovered at the coordination meeting.
Everyone builds from the current revision — and there is a defensible record of who knew what, when.
Last updated
Zepth Core module
#1
cause of construction disputes in North America: errors and omissions in contract documents
Arcadis
44.8%
of projects affected by incorrect, incomplete or late design information — more than scope change itself
HKA CRUX Insight (2,200+ projects)
32.1%
of Middle East projects hit by late design information alone — nearly double the rate elsewhere
HKA CRUX Insight
5.5 hrs
a week spent just looking for project information
FMI research
$88.7B
of global rework in a single year caused directly by bad data — the wrong version, the wrong document
FMI research
The document register is the single authoritative list of every drawing and document on the project: number, title, current revision, status, issue date, and who holds it. Its entire purpose fits in one sentence — everyone builds from the current revision, and there is a defensible record of who knew what, when.
It is the least glamorous instrument on a project and one of the most consequential. Projects that get it right avoid a remarkable share of construction’s most expensive arguments; projects that don’t discover the cost in rework, in late-design claims, and in an as-built set nobody trusts.
The number one cause of construction disputes in North America, per Arcadis’s annual analysis, is errors and omissions in contract documents. HKA’s CRUX research across 2,200-plus projects found incorrect, incomplete or late design information affecting 44.8% of them — more than scope change itself. In the Middle East, late design information alone hit 32.1% of projects, nearly double the rate elsewhere.
Under that headline sit the daily costs: about 5.5 hours a week per professional spent just looking for project information, and an estimated $88.7 billion of global rework in a single year caused directly by bad data — the wrong version, the wrong document, the wrong answer.
Revision control is the core discipline. Each revision supersedes the last; superseded documents are archived — never deleted — with the chain intact, and holders of controlled copies are notified. The classic site failure needs no elaboration: a printed set in the site cabin at Rev B while Rev D sits in the system. A controlled copy is tracked and updated. An uncontrolled copy — the email attachment, the WhatsApp photo of a drawing — is a snapshot nobody will ever update.
Status codes say what a document may be used for. Under ISO 19650’s UK conventions: S0 work-in-progress, which never leaves the authoring team; S1–S5 shared for coordination, information or review; the A-series published and accepted for a stage — suitable for construction, for instance. Alongside this lives the separate consultant review system familiar across the GCC: Code A approved, Code B approved with comments, Code C revise and resubmit. Two different systems answering two different questions — what a document is for, and what a reviewer said — and conflating them causes real site errors. Building from a “for information” issue is a failure mode with a body count of budgets.
The register is delay-claim evidence — in both directions. Under FIDIC-style contracts the engineer must issue drawings in reasonable time, and the register and transmittal log are the contemporaneous proof of planned versus actual release dates when design comes late. A deliverable “issued on time” but returned Code C twice was effectively late — which is why sophisticated registers track approved-for-construction dates, not first issue. And the same register that proves the employer’s late release will just as happily prove the contractor sat on current information for three weeks.
As-builts are a rolling product, not a closeout scramble. What is installed drifts from the IFC drawings through site instructions, RFI answers and field fixes. Capture redlines against the register as work proceeds, or reverse-engineer an “as-built” fiction at handover — with hidden services buried behind walls as the highest-stakes unknowns, and the facilities team paying for the fiction for a decade.
The industry term is “common data environment,” and most practitioners have never used it in a sentence. Here is what it actually means: one shared filing system for the whole project, with rules. Every document has one official current version; a status that says what you are allowed to use it for; and a record of everyone who received it. Documents move from work-in-progress to shared to published through checks — not by someone attaching a file to an email.
That is the whole idea. ISO 19650 standardises the naming, the statuses and the gates. The value is the discipline, not the acronym.
The register is live, not a spreadsheet maintained by heroics. Numbering rules and naming conventions are enforced on upload. Revisions supersede cleanly with full history. Status and review codes are tracked, distribution is recorded, and every document is viewable in the browser — including native CAD and BIM formats — without desktop software.
Bulk upload handles the thousand-drawing tender set. The audit trail handles the questions that arrive two years later.
Nobody builds from a superseded revision — the current one is unambiguous, and the retired one says so.
Late-design release is provable: planned versus actual issue dates, with the transmittal trail behind them.
The as-built set is assembled continuously, not reverse-engineered at handover.
Two years later, the question “who held what, when” has an answer.
Your numbering standard and naming convention applied on upload — the register stays consistent because non-compliant files are caught at the gate.
Clean supersession with full history. The superseded document is archived, not deleted, and its holders are notified.
Suitability status (what a document may be used for) tracked separately from consultant review codes (what a reviewer said).
Every controlled copy has a named holder and a timestamp, issued through transmittals rather than email.
Drawings, specs, native CAD and BIM formats — no desktop software, no “can you export it as a PDF?”.
The thousand-drawing tender set lands in one operation, registered and numbered.
Numbering rules and naming conventions enforced at the gate, so a non-compliant file never becomes a non-compliant register entry.
What the document is for — work-in-progress, shared, published for construction — recorded separately from what a reviewer said about it. The two never get conflated.
Distribution runs through transmittals, so every controlled copy has a named holder and a timestamp. Unacknowledged issues are visible, not assumed.
The new revision retires the old, holders are notified, and the superseded document is archived with the chain intact — never deleted.
Site instructions, RFI answers and field fixes captured against the register while the work happens — so the as-built set is a product, not a reconstruction.
Late deliverables against IFC dates, documents stuck in review, revisions issued but unacknowledged by their holders — surfaced continuously, not discovered at the coordination meeting.
Ask “which drawings changed in the last structural issue, and who is holding superseded copies?” and get the answer with document links, from live project data.
The register connects to everything citing it — the RFI answered against Rev C, the inspection passed on Rev D — so revision impact is visible instead of archaeological.
Planned versus actual issue dates assembled into a late-design-release chronology in minutes, with every transmittal referenced.
The engineer’s judgment stays in charge; the AI removes the latency and the blind spots.
A live register filterable by number, revision, status, discipline and package, with outstanding-revision, overdue-for-issue and unacknowledged-distribution views. Planned versus actual issue dates export as a late-design-release chronology, with the transmittal trail attached.
Minimum viable columns: document number, title, current revision, status or suitability, issue date, purpose of issue, and distribution — who holds controlled copies. Anything less and you cannot answer “who is building from what.”
S-codes (the ISO 19650 convention) describe what a document may be used for: work-in-progress, shared for coordination or review, published for a stage. A/B/C codes are a consultant’s review verdict on a submission: approved, approved with comments, revise and resubmit. One is a usage status, the other a review outcome — a document can be Code B and suitable for construction once the comments are incorporated.
A controlled copy is registered to a holder who receives every update. An uncontrolled copy — an emailed PDF, a printout, a photo — is frozen at the moment it left the system, and nobody will update it. Sites are built from uncontrolled copies more often than anyone admits.
Read the full answerThe document controller, enforcing standards set by the project’s information management plan. On disciplined projects the DC has the authority to reject non-compliant uploads and chase acknowledgments; on undisciplined ones, everyone maintains their own truth.
Yes — it is a core contemporaneous record for late-design-release claims: planned versus actual issue dates, revision histories, and transmittal linkage. It cuts both ways, which is an argument for keeping it honest, not for keeping it vague.
Read the full answerNo. A folder shares files. A common data environment enforces one current version per document, a status that governs its use, gated movement between work-in-progress and published, and a full audit trail. The workflow is the point — the storage is incidental.
Read the full answerRelated modules
Related answers
In the guides
Terms defined here
Zepth is the construction project delivery platform — it runs construction, procurement and asset management on one record, and does the work: reading the drawings, reviewing the submittals, matching the invoices and flagging the risks, with a human sign-off on anything consequential.
A short, tailored walkthrough on your real workflow — no generic demo.