What Happens to Photos in the Review Folder After Culling?

Why frames end up there in the first place

Every frame Cull AI Studio scans gets a confidence value alongside its sharpness, exposure, and face analysis. When that confidence is high enough, the frame is sorted directly into KEEP or REJECT. When it falls below the app's uncertainty bar, the software does not guess; it routes the frame to REVIEW instead. That design choice means the REVIEW folder is not a random sample of your shoot, it is specifically the frames where the AI itself was not sure, which is exactly the set where a human look is most useful.

What you actually do with it

Working the REVIEW folder means going through it and deciding each frame: promote it to KEEP or confirm it as a REJECT. Because these are the genuinely ambiguous frames, expect a mix, some marginal exposure calls that resolve quickly once you look at them, a few near-duplicate bursts where the difference is a subtle expression, and occasionally a story-critical frame that scored poorly on pure technical measures but clearly belongs in the gallery. Giving a reason for your decision when you correct a call, closed eyes, better expression, missed focus, turns your review into more useful training data than a bare approval or rejection.

A concrete example

After a reception-heavy wedding with chaotic lighting, a photographer's REVIEW folder holds a few hundred frames, well beyond the technical rejects and confident keeps sorted automatically. Working through it, most confirm the app's instinct either way once she actually looks; a handful get promoted, including a soft-focus frame of the couple's first dance she decides is worth keeping despite scoring low on sharpness. She logs that correction with a reason, and it becomes part of what her personal model learns about her tolerance for that kind of frame.

What happens to the record afterward

• Your decision is final for that frame. Once you confirm or flip a REVIEW frame, it moves to KEEP or REJECT and is treated the same as any other decision from that point forward.
• The correction is logged. Every REVIEW decision, along with any reason you provide, is recorded in the project's audit trail.
• It feeds training. Once roughly 1,000 reviewed decisions accumulate, they contribute to your personal model, with reversals weighted double.

Honest limitations

The REVIEW folder only surfaces what the AI itself flagged as uncertain; it does not promise that every frame worth a second look ends up there, since a confidently wrong decision elsewhere in KEEP or REJECT would not appear in this queue. Clearing the REVIEW folder also takes real time on a wedding with a lot of ambiguous lighting or expression calls; it is not instantaneous just because it is smaller than the full take. Treat it as the highest-value part of your review, not the entirety of it.

How Cull AI Studio handles this

Cull AI Studio routes any frame below its default 0.65 confidence bar into a REVIEW folder and queue rather than deciding it silently, so your attention goes to the frames where it is most needed. Every confirmation or reversal you make there is recorded with its reason in the project's audit trail, and reversals carry double weight when your personal model trains locally on your computer. For the definition of the underlying concept, see what a review queue is, and for a full walkthrough of reviewing AI output more broadly, see how photographers should review AI selections.