Responsible AIPublic guide
Use AI assistance safely
Request bounded assistance, review provenance and decide whether to apply, revise or reject the result.
Outcome
Request bounded assistance, review provenance and decide whether to apply, revise or reject the result.
QuantifyQS keeps the working record understandable to a quantity surveyor and reconstructable by an authorised reviewer. The screen helps organise the work; the accountable professional still owns judgement and approval.
Before you start
- A permitted task and minimum necessary context.
- Workspace consent and available AI budget.
- A named professional reviewer.
Working method
- Confirm contextOpen the correct organisation and project, confirm your visible role, then check that source information and status are current.
- Make the controlled changeComplete the use ai assistance safely task in the relevant workspace. Use references, notes and structured fields instead of relying on memory or an offline copy.
- Review before issueCheck quantities, rates, dates, parties and evidence, then use the named review or approval state appropriate to the workflow.
- Retain the resultKeep the issued output and its project event together. Where a PDF or spreadsheet is exported, verify the generated file before sending it outside the workspace.
Evidence and checks
- Provider, model, request and outcome are recorded where live.
- Human decision is explicit.
- Applied output is tied to source and project context.
Roles and boundaries
- Workspace role determines what a person may view or change; plan determines what the organisation has contracted to use.
- Customer-specific Firm and Enterprise panels do not widen a user role or cross the active tenant boundary.
- AI suggestions, imported data and generated documents require human review before professional reliance or issue.
Did this guide help?
Only your answer, guide slug and version are recorded.