AI Test Sessions

/casewriter/ai-tests lists every AI generation session in the active project.

Session Metrics

Four cards sit above the list.

Card	Shows
Total Sessions	All sessions, with the number completed
Cases Generated	Total cases, with the average per session
Running Sessions	Sessions in progress, with the number pending
Avg Generation Time	Average duration, with the completion rate

Status Tabs

The list is split by session status, each tab carrying its own count.

Tab	Meaning
All Sessions	Everything
Pending	Queued, not started
Running	Generating now
Completed	Finished normally
Needs Review	Finished, but the results want a look before use
Stopped	Stopped by a user
Generation Failed	The run failed; the card shows the reason
Cancelled	Cancelled

Finding a Session

• Search matches the session name, code, source file name, or ID.
• Sort by date newest or oldest, name, status, or case count.
• Three facet filters narrow by Execution, Focus, and Level, matching the session's configuration.

When filters leave nothing on screen, the empty state offers Clear all, which also returns the tab to All Sessions.

Session Cards

Each card shows the session name, its code, cases generated, duration, and the configuration chips for execution type, focus, and level. A status badge and timestamp sit at the bottom; a failed session also shows its error reason.

The card menu offers:

Action	Notes
Open	Opens the session detail page
Copy URL	Copies a shareable link to the session
Rename	Requires the review permission
Delete	Requires the generation delete permission

Deleting a session removes the session and its generated test cases, and the metrics above the list drop accordingly.

Session Detail

Opening a session shows everything about the run at /casewriter/ai-tests/<id>.

Header

The title is the session name, with the creation date, source file name, and duration underneath. On the right sit the status badge and the available controls.

• Stop appears while the session is running.
• Re-Generate appears on a finished session that was not generated from a mindmap, and reopens the generator prefilled with this session.
• A session that carries a mindmap offers View Mindmap instead, which reopens the map so you can edit it and start a new run from it.

Running Sessions

A progress card reports the current phase, percentage, cases produced against the target, elapsed time, and an estimated time remaining. The page polls while the session is active and stops once it reaches a final status.

Banners

• A failed session shows the backend's reason for the failure.
• A Needs review session says how many of the requested cases were produced and asks you to review them before use.

Context and Configuration

A card lists the AI Model, Language, Target Count, and Platforms used for the run, plus a Testing Configuration block with execution type, focus, level, and the selected test types.

Input Sources

A summary strip shows every source the run used — documents, requirement text with its character count, URLs, PBIs, images, a tracker issue, or a SmartRequest.

Selecting it opens a dialog with the detail:

• Documents, with a download button for each file
• The full requirement text
• URLs as clickable links
• The PBI list, with the source PBI session
• Image thumbnails that open a full preview
• The linked issue's key, summary, tracker, and priority
• The SmartRequest's title, description, status, and included data

Generation Progress

A second card shows the cases produced against the target and the run duration, with a progress bar.

• If the run produced nothing, a warning explains that it finished without producing any cases and suggests checking the inputs and model configuration.
• If it produced fewer cases than the target, a quiet note explains that duplicates were left out to keep the set distinct.

Results

Below the cards, the full results list is the same surface described in AI Test Case Generation — search, sort, grouping, column filters, expandable cases, regeneration, review actions, Excel export, and Save to Repository.

Deep Links

Link	Effect
?expandRegenerated=true	Opens the session with every case expanded
?operationId=<id>	Opens the regeneration comparison dialog for that operation
?code=<case code>	Expands and scrolls to a single generated case

Billing transactions and issue detail screens in other modules link into sessions this way.

Related

• Review Workflow — approving what a session produced
• AI Test Case Generation — starting a new session
• Test Repository — where approved cases land