Operate and Review Scout Tasks
Scout keeps scope, source evidence, recommendations and decisions together throughout a task. Each review should identify the current stage, resolve the first blocker and leave a clear handoff for the next responsible person.
Review Task Inputs
- Confirm the tenant, task name, equipment or dataset scope.
- Review the evidence categories, time range and selected sources.
- Check the latest completed step and the current blocker.
- Look for newer source, model or business-requirement versions.
- Assign the next owner and record the information they need to provide.
Assign the Next Action
| Current state | Owner | Next action |
|---|---|---|
| Source cannot be read | Source owner or deployment engineer | Check access, connectivity and an actual observation |
| Equipment is missing or duplicated | Asset or MDM steward | Reconcile the master-data identity and review queue |
| Field meaning needs confirmation | Domain specialist | Provide a field dictionary, manual or installation record |
| AI asks for clarification | Task owner and relevant specialist | Add facts and references, then resume analysis |
| Candidate is ready for review | Reviewer | Confirm, reject or request evidence against the source record |
| Validation is blocked | Data operator | Correct the named gap and rerun the affected step |
| Release awaits approval | Approver | Review the exact versions and change preview |
| Release is approved | Release operator | Apply the approved plan and inspect its result |
| Published result lacks readings or business output | Data or application owner | Trace the source observation and downstream processing path |
Review Candidates and Changes
For equipment field associations, verify:
- equipment identity and attributes
- measurement quantity, unit, scaling and installation location
- real-time or historical source role
- sample time and source freshness
- supporting evidence, conflicting evidence and open questions
For business datasets, also review identity keys, null rules, relationships, duplicate records, rejected rows and conflicts. For model changes, compare the active model with the proposed draft. See Brick Models and Data Association.
Validate, Approve and Apply
- Complete the decisions required by the selected workflow.
- Run sample validation, a dry run or the release preview.
- Confirm the evidence still matches the current source, model and binding versions.
- Have the approver review that exact change plan.
- Have the release operator apply the approved plan.
- Record the effective version and operation result.
Dataset and semantic releases separate approval from execution. When a version changes, regenerate validation evidence and approval. If a request times out, refresh the task and inspect the existing operation before retrying.
Verify the Outcome
| Workflow | Result to retain |
|---|---|
| Data exploration | Sources inspected, fields confirmed, open questions and scope |
| Equipment business check | Approved bindings, actual readings, observation time and application-use record |
| Dataset release | Effective dataset version, MDM and fusion results, downstream check |
| Brick semantic release | Effective model version, data references and source observations |
| Historical CSV preview | Mapping preview for the current file and owner of the governed import |
Record missing data, currently unavailable data and an observed value of zero as distinct results. When an application already shows a chart, also verify its data version and source.
Handover and Revisit
Every handover should include the latest completed step, current blocker, owner, required evidence and next review time. Revisit affected tasks when a source, model or business requirement changes. For recurring business data, follow Repeat Dataset Deliveries.