Sessions
Sessions is the operational center for investigating AI sessions across the tenant. Use it to find the exact set of sessions you need for a case review, export, escalation, or audit response.
It supports direct lookup, full-text search, and filtering by application, model, outcome, risk, intent, policy score, review state, and blockchain anchor status.
Sessions
The Sessions workspace is the main place for finding and reviewing AI sessions across your tenant. Every ingested session is searchable here by risk, outcome, model, intent, time window, and review state.
All roles reach the Sessions page. What you see inside each session depends on your role. Administrators, Developers, and Governance Engineers see the full step-by-step execution trace and raw governance attributes. Compliance Officers see session summaries and policy outcomes — not the raw trace steps. Business Owners see session summaries with sensitive content masked. Auditors see sessions scoped to their active engagement, including the full trace, but with sensitive content masked.
You see the full step-by-step execution trace, raw governance attributes, and sensitive field values. PII fields remain masked — only Administrators and Compliance Officers can reveal PII. Use the trace to confirm your SDK is sending the correct span structure before you promote to production.
You see the full step-by-step execution trace and governance attributes — the same detail level as Administrators and Developers. PII fields remain masked in your view. Use sessions to verify that your active rules are evaluating the expected signals and to identify representative sessions worth promoting to your evaluation datasets.
You see the session summary, policy outcome, and risk classification — not the raw execution trace. Sensitive captured content and PII fields are masked in your view. Use the filters to find sessions that need escalation or approval.
You can bulk sign-off and archive sessions, export subject data, and request data erasure from the session detail panel. You see policy outcomes and session metadata — not the raw trace steps.
Your session list is scoped to your active audit engagement. You see the full step-by-step trace for in-scope sessions, but sensitive captured content is masked in your view. You can export evidence packs from the session detail and annotate sessions to capture your audit findings. You cannot bulk sign-off sessions, modify session review status, or alter session data.
Finding Sessions
The Explorer shows the newest sessions first. Use the filter panel on the left to narrow the list.
Filter Dimensions
| Filter | Options |
|---|---|
| Date range | Last 24 h, 7 days, 30 days, 90 days, or a custom range |
| Application | Limit results to one or more specific applications |
| Risk level | LOW, MEDIUM, HIGH, CRITICAL (multi-select) |
| Outcome | Success, Partial Success, Failed, Error, Cancelled, Timeout (multi-select) |
| Guardrail action | Blocked, Flagged, Allowed, None |
| Model | Any model identifier — type to search |
| Intent | Any intent label — type to search |
| Policy score range | Minimum and/or maximum score |
| Human oversight | Sessions with or without human oversight annotation |
| Blockchain anchored | Confirmed anchored / anchoring pending / not anchored |
| Review status | Pending review, Reviewed, Needs escalation |
| Session ID | Direct lookup by VeriProof session ID or your own trace ID |
Filters use AND logic across categories. To broaden within one category, select multiple values in that filter.
Search by ID
To navigate directly to a specific session, paste any of the following into the search bar at the top of the Explorer:
- VeriProof session ID (UUID format)
- Your application’s own trace ID (if passed as
externalIdduring ingest) - First 8 characters of either ID
Reading the Session List
Each row in the Explorer shows:
| Column | Description |
|---|---|
| Risk level | Color-coded badge: blue (LOW), yellow (MEDIUM), orange (HIGH), red (CRITICAL) |
| Session ID | Truncated, click to copy the full ID |
| Application | The application that ingested this session |
| Outcome | Success / Failure / Escalated / Cancelled |
| Policy score | 0–100 score for this session |
| Model | Model identifier |
| Timestamp | When the session was ingested |
| Review | Review status badge — Pending, Reviewed, or none |
| Anchor | Blockchain anchor confirmation status |
Click any row to open that session in Time Machine for full forensic detail.
Bulk Review
The Explorer supports bulk review when you need to clear or escalate a large cohort of sessions, such as all HIGH-risk sessions from a quarter-end review.
Bulk review is available to Administrators, Governance Engineers, and Compliance Officers only. Bulk actions are permanent and appear in the policy audit log with your identity.
To perform a bulk review:
- Apply filters to select the cohort of sessions you want to review.
- Click Select all (or select individual sessions using the row checkboxes).
- Click Bulk actions → Mark reviewed (or Escalate or Flag for attention).
- Add an optional note that will be attached to all selected sessions.
- Confirm. The action is applied immediately.
Bulk review is limited to 500 sessions per operation. For larger cohorts, export first and use the Evidence Export workflow.
Exporting Sessions
To export the current filtered set of sessions:
- Apply your filters.
- Click Export → CSV or Export → JSON.
- The export includes the fields visible in the list view plus policy and annotation fields.
For tamper-evident audit packaging, use the Evidence Export workflow from the application workspace instead.