Why we show the snap twice

BoardSnap shows your whiteboard snap in two places:

1. Immediately after capture — a full-screen preview that you review and confirm before analysis begins.
2. At the top of the summary card — a thumbnail of the board alongside the summary text.

Most photo-analysis apps show the image once, maybe as a thumbnail in a list. Showing it twice is a deliberate choice. Here's why.

### The first show: verification

After the shutter fires, you see a full-screen preview of the processed image — VisionKit's perspective-corrected, cropped version of what you just photographed.

This is the verification step. It gives you the chance to confirm that the image is usable before committing to the analysis. If the glare ate half the board, you'll see it now. If the crop is wrong, you'll see it now. One tap to retake, one tap to confirm and proceed.

Without this step, we'd be sending images directly to analysis — and users would only discover a bad snap when they got a bad or incomplete summary. That's a worse experience than discovering the bad snap immediately and being able to retake it in seconds.

The verification step adds about 3 seconds to the flow in exchange for eliminating a whole class of "the AI read it wrong" support requests. Worth it.

### The second show: grounding

The thumbnail at the top of the summary card is there for a different reason: grounding.

AI-generated text can feel unmoored, especially in a productivity context. "Did it actually understand the board? Is this summary connected to what was actually on the board?"

Showing the board thumbnail alongside the summary keeps the summary grounded in the physical artifact it came from. The connection between the whiteboard in the room and the text on the screen is visible and explicit. Users don't have to hold the connection in their heads — they can see it.

This grounding effect is subtle but measurable: in the version of the app that didn't show the thumbnail alongside the summary, users were more likely to manually verify the summary against the original photo (checking their camera roll, going back to the board). With the thumbnail visible, that behavior dropped significantly. The visual anchor reduced the verification anxiety.

### The double-show as a trust mechanism

Both instances of the snap serve trust purposes:

The first show: "You're in control. We're not sending a bad image to AI without your approval."

The second show: "The summary is grounded in your actual board, not in AI confabulation."

Trust is expensive in AI products. Every moment of "did it hallucinate that?" or "is this actually based on what I snapped?" is an erosion. The double-show is cheap infrastructure for preventing those erosions.

### The exception: the list view

In the board list view, we show a thumbnail for each board — which means users see their board snap a third time, as a visual identifier in the list.

This is different from the grounding function. In the list view, the thumbnail is a visual anchor for navigation: you recognize your board by its image. The summary text alone isn't a reliable way to find a specific board if you have many. The image is faster.

Three places: capture confirmation, summary grounding, list navigation. Each is doing different work.