BIM common data environment failure on a mixed-use tower

I would separate the technology from the process and the contractual allocation. The CDE may be hosted by the contractor, but the failure is that information containers were not controlled through defined states, suitability codes and approval routes. Weekly folder uploads with names like 'final' and 'final2' show there is no authoritative version; overwriting a structural model shows there is no audit trail or permission model that protects published information. The contractor building from a superseded grid indicates that the exchange process did not distinguish work-in-progress from information suitable for construction. I would trace the specific grid, the model version, the transmittal record, the RFI history and the decision log. The project director's statement that the CDE is the contractor's system is not a sufficient answer: the lead designer still has information-management obligations, and the practice may have issued or accepted information without checking its status. The root-cause report should identify process, governance and contractual causes, not just name the person who uploaded the wrong file.

I would implement a single CDE with role-based permissions and a controlled workflow. Every information container should have a unique identifier, discipline code, originator, revision and status or suitability code. Work-in-progress should be separated from shared, published and archived information. Only authorised task-team leads should be able to promote information to a shared or published state, and published information should be immutable with a full audit trail. Transmittals should record what was issued, to whom, when and for what purpose. The federation process should run on a fixed cycle with clash reports and model-health checks before issue. Naming conventions should be enforced by the system, not by email reminders. I would also define the information requirements at the outset, including level of information need, delivery dates and acceptance criteria. The 620 open RFIs suggest the workflow is not resolving queries fast enough, so I would add a query status, owner and due date, and report on overdue items at each project meeting.

I would advise that the claim turns on causation and contemporaneous records, not on a general assertion that the CDE was messy. The client should preserve the model history, transmittals, RFIs, meeting minutes, email and the contractor's build records. The key questions are: which information was the contractor entitled to rely on, when was the superseded grid published, was the contractor told about the revision, and what instruction or delay followed? The overwritten model is serious, but the contractor may also have failed to check the status code or to raise an RFI before building. I would produce a chronology and an information-quality assessment, and separate technical causation from contractual entitlement. I would also check whether the practice's own quality reviews and transmittal approvals were adequate, because the client will want to know whether the lead designer contributed to the delay. I would not offer a legal conclusion; I would give the client a factual and technical basis for its lawyers to assess the HKD 28 million claim.