A construction project’s exposure extends across consultants, contractors, cloud services and connected equipment. Resilience depends on knowing which activities rely on those connections and how work can continue safely when they fail
An inaccessible project platform creates questions that reach well beyond the IT helpdesk. Which drawings can the site team rely on? Can inspections proceed? Who can authorise a change, and how will the resulting records be reconciled when service returns? A cyber incident may affect the confidentiality, accuracy, or availability of information. Those consequences demand different responses. A file that opens successfully can still be untrustworthy if an attacker or compromised account has changed its contents.
Construction teams should therefore define cyber resilience around the decisions and activities that digital systems support. The relevant unit is often the project, whose information crosses several organisations, rather than a single company’s network. Strong controls at the main contractor cannot, by themselves, settle how a consultant shares files, how a subcontractor’s staff use accounts, or how a specialist maintains remote access to installed equipment. Each connection needs a clear purpose, a responsible owner, and an agreed response if it becomes unsafe.
Map dependencies before buying reassurance
The National Cyber Security Centre’s supply chain assessment guidance is aimed at procurement, risk and security teams seeking confidence in their suppliers. Its staged approach provides a useful foundation for examining construction relationships. A project dependency map should show who provides each key service, what information they hold or change, and what other services they depend on. This can reveal a shared point of failure even when the project appears to use several independent suppliers.
Start with decisions whose interruption or corruption would have serious consequences. Approved design information, payment authorisations, site access and building controls may need different treatment from routine meeting documents. Record the maximum interruption each activity can tolerate and the people who determine whether work can continue. These are operational judgements, not numbers for a software vendor to choose alone. The exercise also makes procurement more specific: assurance should address the actual service and its dependencies, rather than rely on an impressive but unrelated certificate.
Give access an owner and an expiry
The NCSC’s guidance for small and medium-sized organisations brings together practical protection, incident planning and recovery resources. Its relevance extends to larger construction projects that depend on smaller firms. Basic measures such as supported software, secure accounts and resilient backups need to work at organisational boundaries. A project cannot assume that every participant has a dedicated security team, so onboarding should explain required controls in language a small business can act on.
Within a common data environment, named accounts make it easier to connect actions to authorised users. Permissions should follow the task: viewing a published drawing, uploading a proposal and approving its release are different activities. A temporary specialist should not retain broad access after their package finishes. Build joiner, mover and leaver checks into project administration, including people who change employer while remaining on the same scheme. Multifactor authentication and controlled privileged access need practical support so that inconvenience does not lead users to share credentials or move work into unmanaged channels.
Protect the integrity of released information
Availability attracts attention because an outage is obvious. Integrity can be harder to recognise. If a compromised user replaces a drawing, alters a payment destination or changes an equipment setting, people may continue working with apparently normal information. Project workflows should therefore separate preparation, checking and authorisation where consequences justify it. The evidence of approval should remain connected to the specific version, with a usable history of later changes. Merely showing that someone clicked an approval button does not explain what they approved.
Payment instructions deserve an independent verification route, particularly when bank details change. For technical information, identify how teams will establish which release was authoritative if an account or platform is compromised. Logs can help, provided they are retained, protected and available to the people investigating. A recovery plan that restores files without checking their status may reintroduce manipulated information. The project director and technical leads need to agree how confidence will be re-established before dependent activities resume, using proportionate checks for the affected work.
Treat connected assets as operational systems
Remote access to a building management system is different from access to a document library. Changes can affect physical operation, and maintenance windows may be constrained by building use. The NCSC’s operational technology guidance addresses systems that monitor or control physical processes. Its collection includes secure connectivity principles. For a connected building, security and engineering teams need to work together so that network changes, patching and emergency responses account for the consequences of interrupting the equipment.
At procurement, establish who owns administrative accounts, how remote support is authorised, which components require external connectivity and how long the supplier will support them. At handover, remove temporary commissioning arrangements and verify the intended operating configuration. A specialist’s remote connection should have a documented purpose and a route for revocation. Asset owners also need to know what happens if the supplier ceases trading or the connected service ends. These questions are easier to resolve before installation than after an unsupported controller becomes embedded in a critical building service.
Limit personal data and prepare the reporting route
Construction records can include identifiable workers, residents, visitors and members of the public. Photographs, access logs, incident reports, and contact lists can contain personal data even when the system’s main purpose is technical. Give each collection a defined purpose and retention approach, and restrict access according to that purpose. Information useful for investigating an incident may not need to be visible to every project participant. Copying an entire dataset into a new analytics service can change the exposure without improving the underlying work.
The Information Commissioner’s Office’s personal data breach guidance explains the reporting threshold. A controller must notify the ICO without undue delay and, where feasible, within 72 hours of becoming aware of the breach, unless it is unlikely to create a risk to individuals’ rights and freedoms. High-risk breaches can also require notification to affected people. Not every cyber incident is reportable, but assessment and recording are necessary. Project arrangements should make it clear who investigates, who informs the relevant controller and who decides on notifications.
Rehearse recovery with the people doing the work
A practical exercise can start with a hypothetical loss of the common data environment during a critical stage. Ask the site manager, design lead, commercial team and key subcontractor what they would do next. Can they obtain verified information? Which activities must pause? How will inspections and instructions be recorded without creating conflicting histories? The aim is to expose dependencies and unclear decisions while there is still time to improve them. An exercise should use safe test data and avoid disrupting live building or site controls.
Then test restoration. Backups are valuable only if the organisation can recover the necessary information, permissions and context within an acceptable period. Establish how much recent work could be lost and how it would be reconstructed. A cloud service’s general recovery capability may differ from the recovery available to an individual customer after accidental deletion or account compromise. Record the result of an actual recovery test, the unresolved gaps and the person responsible for correcting them. This is stronger evidence than a contractual assurance that backups exist.
Make resilience proportionate and accountable
Smaller suppliers need achievable requirements and a clear support route. A single questionnaire applied to every organisation can produce paperwork without revealing the risks that matter. Tailor checks to access and consequence, while maintaining essential controls across the project. Where a supplier cannot meet a requirement immediately, decide whether the work can be redesigned to reduce exposure, whether additional support is justified or whether access must wait. The decision should belong to an identified risk owner, with a review date rather than an indefinite exception.
