Timeline
The timeline is a case’s complete activity log. Every meaningful change is recorded in order: the moment the case was created, every piece of evidence added, every hypothesis updated, every link drawn on the board, and the conclusion that closed it. It is the case’s memory and its audit trail.
Open it from the Timeline panel on the right of the case board. The panel has three tabs:
| Tab | What it holds |
|---|---|
| Activity | Everything done on the case, by people and by Autopilot, grouped by day |
| Chronology | The story of what happened in the real world, in the order it happened. You build it |
| Threads | Every hypothesis and discussion, with Show on board or Place on board |
What Activity records
| Area | Events you’ll see |
|---|---|
| Case | Opened; title, description, severity or status edited (one entry while someone keeps editing, with the old and new values); conclusion written or changed; clean-up settings and finding filters changed |
| Inquiries | Linked, unlinked, auto-add switched on or off, and findings pulled in from their matches (by a person, or added by the watch itself), each named, with what the filters kept out. From each linked watch’s own history: what a scan changed in its answers, new ones and ones gone |
| Evidence | Assets added or removed, findings attached or removed, notes on them edited, and what clean-up and filters took out, and why. Leads accepted or dismissed (with the reason), and new leads the case found: one entry per stretch of automatic refreshes, saying how many of each kind |
| Escalations | Escalation rules added, changed or removed; findings that escalated (which rule, and what brought them in); escalations cleared. Escalations carry a warning sign and have their own filter |
| Hypotheses & threads | Threads started, notes added, statements revised, verdicts and confidence changed, evidence linked or unlinked. Hypothesis rules added, changed or removed; answers a watch linked to a hypothesis by itself (one entry per watch while its scans keep landing, with how many findings went to which hypothesis and with which stance); a hypothesis deleted together with its evidence |
| Board | Notes and frames added, edited or removed; links drawn, edited, removed or made global; highlights; tidy-ups; snapshots; comments resolved |
| Autopilot | Each Autopilot run on the case |
Reading the feed
The feed is built to scan quickly:
- Grouped by day, with clear date headers, so you can follow the investigation as a story.
- Each event shows who did it: a teammate, or an Autopilot agent, clearly badged. Human and AI work sit side by side in one trustworthy record.
- Filter by area to see only evidence, only hypotheses, only the board, and so on.
- Show on board: events about something still on the board take you straight to it.
- Link to an entry: the link icon on an entry copies its address. Opening it opens the board with the timeline scrolled to that entry and marked. Warnings elsewhere on the board link to their entry the same way, with What changed: evidence gone from its source, answers a watch lost, what a clean-up or a filter took out.
- What the platform does by itself reads as one entry. A watch’s auto-add and the automatic clean-up fold every pass into one growing entry while they keep happening (a pass within the hour since the last): Auto-added by a watch, 37 findings over 6 scans of Firmenbuch between 14:02 and 15:40. The entry lists the findings, newest first, and moves up to the time of its latest pass. What a person does is always its own entry.
Chronology
Activity records what was done to the case. Chronology is different: it is the sequence of real-world events the investigation uncovers. For example: 12 March, 23:40: payroll export emailed to a private address, then 13 March: mailbox reported.
- Add an event with a title and the date and time it happened.
- Events are shown in the order they happened, not the order they were recorded.
- Events proposed by Autopilot are marked as such. Verify one once you have confirmed it; verified events are marked too.
- Edit or remove events as the picture changes.
A good chronology is often the clearest summary of a case: exactly what happened, when, and how sure you are of each step.
Because both people and Autopilot act on the same case, the timeline is where you confirm exactly what the AI did and when. It follows the same “explain everything” principle as the Autopilot flight recorder.