Good morning, Alderline.
Here’s the work asking for your team’s attention today.
AI cannot approve its own proposal.
Customer messages and external changes stay blocked until a person reviews them.
Operating pulse
What needs a decision, an owner, or a clear next step.
Team queue
Small, concrete next steps.
Decisions with evidence
Keep the original customer context attached to each choice.
Workspace notes
Operating context for this sample company.
Alderline is a fictional 10-person product studio. Its sample activity is written as a believable history; it does not represent customer claims or connected account data.
Customer signals
Review incoming requests without losing the source, timing, or uncertainty behind them.
| Request | Source | Interest | Status | Owner | Last activity |
|---|
Sample history is synthetic. Selecting a row opens its evidence and review controls.
Preview a feedback migration.
Start from an export you control. See which history maps cleanly and what needs separate handling before you consider moving anything.
1 · Export from your current tool
Use the source product’s own CSV export.
2 · Choose a CSV file
Maximum 2 MB · preview stays on this device
The dry run detects missing titles and repeated titles. It does not merge, overwrite, or import records.
Dry-run preview
Field mapping
What this export does not carry
| Source title | Vote count | Status | Import check |
|---|
Projects & work
Move approved decisions into owned work, with a simple view of progress and blockers.
Connections
Choose the systems VTB Operation may read. Every integration starts with a narrow, reviewable permission set.
For engineers · event contract preview
A concrete data shape to review before a live connector exists.
Example feedback event
{
"event": "feedback.received",
"event_id": "evt_sample_001",
"occurred_at": "2026-11-03T10:26:00Z",
"source": { "kind": "portal", "record_id": "sample_184" },
"content": { "title": "Connect Linear projects", "body": "We copy this every Friday." },
"evidence": { "url": "https://example.invalid/sample/184", "votes": 42 },
"provenance": { "synthetic": true }
}Before a live key is issued
- Publish the API schema and versioning policy.
- Define read-only scopes and tenant isolation.
- Document webhook signing, retries, and deletion behavior.
- Provide a test workspace and revocable sandbox credentials.
A request for these items is a real founder follow-up; it does not create or issue a key.
Six separate requests from Slack, Outlook, and the customer portal are grouped into “Connect Linear projects.” Every source stays intact; the agent only suggests an internal theme for a person to review.
Agent review
Automation can collect, organize, and prepare. A person controls every consequential change.
Review queue
Newest first · proposals are not actions
Proposal detail
Inspect evidence, scope, confidence, and policy before deciding.
Analytics
Understand how requests move through review and into shipped work. Sample metrics are illustrative.
Eight requests have no decision or defer reason. A short review can clarify the next step before more signals arrive.
Choose what the AI can do.
Start with selected, read-only sources and suggestions. Keep customer messages and sensitive changes behind your team’s explicit approval.
What the agent may do
Each switch represents one clear permission boundary.
Decision history
Sample change log · available for team oversight.
Permission levels
Three plain rules, with a clear human checkpoint as impact increases.
Look at only the sources your team selects; no changes are made.
Organize information and draft internal work. A person reviews before it changes the shared workspace.
Messages, publishing, billing, deletion, and access changes stay off or require a separate human decision.
Workspace settings
Manage the workspace profile, team access, notifications, and sample data preferences.