flowchart TD O["Organizations"] --> C["Campaigns"] C -->|"New campaign,<br/>when needed"| I["Connect storage and<br/>index evidence"] C --> E["Evidence"] I --> E E --> N["Notes"] E -. "Related view" .-> V["3D Overview"] A["Kav AI Assistant"] -. "Visible campaign<br/>context" .-> C A -.-> E A -.-> N
17 Data Explorer workflow
Prepare campaign evidence for Integrity review
Data Explorer is where a user organizes inspection data into a campaign, verifies its evidence and context, and prepares it for an Integrity Engineer. A campaign is the working boundary: it brings together organization, inspection evidence, notes, and related spatial views.
17.1 Typical workflow
The flow gives a useful starting point; it is not a lockstep sequence. For example, a user can return from 3D Overview to Evidence, move from Notes to related campaign evidence, or use the Assistant with the current view’s scope intact.
17.2 Workspace surfaces
| Surface | User question | Role in the workflow |
|---|---|---|
| Organizations | Where does this work belong? | Select the organization whose campaign catalog to review. The directory owns organization selection. |
| Campaigns | Which campaign is being prepared? | Browse campaign cards, continue an existing campaign, or start a new campaign. |
| New campaign | Can KAP access and index the source? | A Campaigns action—not a primary navigation stage—for connecting storage, checking access, selecting files, and starting indexing. |
| Evidence | Is the campaign evidence ready to inspect? | Review campaign-scoped annotated evidence, coverage, and related data. |
| 3D Overview | What is the spatial context? | A related map, gallery, and 3D view rather than the default campaign workflow stage. |
| Notes | What observations support review? | Curate campaign observations and supporting files. |
| Kav AI Assistant | What can I learn about this context? | Available throughout the workspace with visible organization and campaign grounding. |
17.3 Prepare a campaign
17.3.1 Establish the working scope
Choose an organization, then browse its campaigns. Campaign cards own campaign selection; the campaign catalog does not duplicate that choice with a second selector. Campaign-required views retain a valid campaign in the URL so evidence, notes, related views, and workspace changes remain grounded in the same work.
17.3.2 Create or continue the campaign
For a new campaign, connect storage, verify access, select the source path, and start indexing. KAP indexes metadata and thumbnails while the raw files remain in connected storage. For an existing campaign, begin directly in Evidence or use 3D Overview when spatial context is the next question.
17.3.3 Review evidence and context
Use Evidence to inspect the campaign’s annotated evidence and coverage. Use 3D Overview to explore related map, gallery, and 3D context without leaving the campaign boundary. Capture observations and supporting material in Notes so the review has relevant evidence rather than an unstructured collection of files.
17.3.4 Ask in context
The Kav AI Assistant is one shared, scope-aware surface. It can be invoked from workspace chrome, the mobile launcher, the sidebar, or a compatible /ask link. Its panel shows the organization and campaign it is helping with; it does not add another page-local organization or campaign selector.
17.4 Handoff to Integrity Engineer
| State | What happens |
|---|---|
| Current | A campaign that exists under an organization is visible in Integrity Engineer. Switching workspaces retains compatible organization and campaign scope. |
| Target | Mark ready for Integrity review will create an explicit evidence-package handoff with readiness status, evidence summary, unresolved exceptions, and accountable ownership. |
The current implicit handoff is useful for exploration, but it is not a governed readiness decision. Until the target state is implemented, campaign presence must not be described as a completed engineering review or an approval.
17.5 Delivery boundary
Q3 sensor-native analysis supports human review of independent RGB and thermal image candidates and broader evidence preparation. Automated cross-modal detection, calibrated evidence-confidence gates, and quantified inspection prioritization remain Q4 targets. The PRD v3.8 defines the delivery status and safety boundaries.