Quality Control Module

Quality Control is the first module in the Batch Manager workflow. The Quality Control module provides operators a focused interface for evaluating image quality and resolving issues flagged by AI scoring or surfaced during operator review. Quality Control opens a batch and presents each image with its AI Quality Score, AI Tags, and user verification state, alongside the image viewer and the tiles needed for QC decisions.

Note: See also the Glossary of Terms for box and batch terminology, and Section 6 Review for image viewer and image editor features shared between Review and Quality Control.

Opening a batch in Quality Control

Quality Control opens from Batch Manager. An operator selects a batch in the Quality Control stage and clicks Play to open the batch in the Quality Control module. Only one operator can work on a batch at a time. Quality Control tracks total time spent on the batch, which contributes to production run rate calculations.

The Quality Control workspace replaces the Batch Manager view while the batch is open. The left side shows a navigation panel listing images in the batch. The center shows the currently selected image in the image viewer. The right side shows the Right Panel Tiles for the current image.

Action buttons

The action button row at the top of the Quality Control workspace provides controls for the current image and the QC session as a whole.

QC Mode indicates the current operating mode and remains active while the operator works in Quality Control.

Enable Image Popup (Ctrl+0) toggles a popup preview of the current image. The popup displays in a separate floating window so the operator can examine the image at a larger size without losing the navigation panel context.

Delete Image (Ctrl+1) removes the current image from the batch. Use this for unrecoverable images that cannot be rescanned. Deleted images appear in the Omitted Images tile and can be restored from there.

Image (Ctrl+2) selects the image view as the main display in the center panel. This is the default view when a batch opens.

Timeline (Ctrl+3) switches the center panel to a timeline view of the batch's QC history. The timeline shows each transition between QC and QC In Progress states with the timestamp, the user who made the transition, and the duration spent in each state. Total time at the bottom of the timeline aggregates all QC time for the batch across sessions.

Quality (Ctrl+4) displays the current quality percentage for the batch. The percentage is calculated against a 100% perfect image baseline based on the number of AI Quality Tags present across the images requiring review. The set of images requiring review depends on the workspace's configured review mode (see Workflow review modes below). The indicator updates live as the operator works through the batch.

AI Image Quality Analysis (Ctrl+5) opens the detailed AI scoring breakdown for the current image. The breakdown includes the Quality Score with its plain-language rating and the timestamp the image was analyzed; a Summary panel with a Photo / Scan Summary and a Document Content Summary explaining the overall quality finding in plain language; a Photo / Scan Issues panel listing each scan-related issue with its severity, location on the page, reasoning, and recommended action; and a Document Content Issues panel listing each content-related issue using the same severity, location, reasoning, and recommended-action structure. The AI Image Quality Analysis is described in detail under Configuration Management (Section 10.2.3 Artificial Intelligence: Image Quality Analysis, Data Extraction, and Redaction).

Properties opens a side panel on the right showing detailed metadata for the currently selected image: System Attributes (Image Name, Session Name, Session Id, Box Id, Box Identifier, Box Image Barcode, Image Id, Image Type, Workflow QC Status, Has Been Rotated, Has Been Cropped), AI Quality Score with the plain-language rating, AI Tags, and the batch timeline. Workflow QC Status holds one of two values: Accepted or Rejected. Rejected images require a replacement and are resolved in the Rescan Module (see Completing Quality Control below).

The image controls row below the action buttons provides image-level adjustments: Show Image toggle, rotation degree, zoom percentage, and Save Rotation. These controls behave the same as in Section 6 Review.

Open Image Editor in the top-right corner launches the same image editor used in Review for cropping and other image manipulation tasks.

Navigation panel views

The navigation panel on the left side has three view modes. The view mode controls appear at the top of the panel.

Thumbnail view (Ctrl+E) shows each image as a thumbnail with filename, AI Score percentage, an X marker for images flagged with AI exceptions, and a User status indicator (a green checkmark when the operator has resolved the image, or a dash when unresolved). Thumb size is adjustable through a slider above the thumbnails.

List view shows each image as a compact row with filename, AI status (a green checkmark for images that passed AI scoring or a red X for exceptions), User status (green checkmark when resolved, dash when unresolved), and AI Score percentage. List view fits more images on screen than thumbnail view and is useful for larger batches.

Exceptions view filters the navigation panel to show only images with AI exceptions — images where the AI score did not pass the threshold or AI tagging surfaced an issue. Use this view to focus on items requiring operator attention. Images already resolved by the operator do not display in this view.

Right-click menu

Right-clicking an image in the navigation panel opens a context menu. Most menu items act on the selected image; Suspend and Complete act on the entire batch.

Image-level actions:

Accept (Enter) marks the image as approved and sets its Workflow QC Status to Accepted. The User status updates to a green checkmark, the batch Quality percentage adjusts, and the navigation panel advances to the next image in the batch. Pressing Enter is the fastest way to work through a queue of flagged images.

Reject (Shift+Enter) marks the image as rejected and sets its Workflow QC Status to Rejected. Rejected images are resolved in the Rescan Module after the QC batch is completed.

Reset (Ctrl+7) clears the operator's prior Accept or Reject decision on the image and returns it to its original AI-flagged or unresolved state.

Rescan / Replace initiates a rescan or replacement flow for the image without leaving QC. Use this when the operator has the original document available and can rescan or replace the image during the QC session itself, rather than routing through the Rescan Module later.

Insert adds a new image before or after the current image position in the batch. Use this when an operator missed a file during the original capture and needs to insert the missing image into the correct position.

Delete (Ctrl+1) removes the image from the batch. Same effect as the Delete Image button in the action button row. Deleted images appear in the Omitted Images tile and can be restored.

Batch-level actions:

Suspend (Ctrl+8) suspends the batch and returns the operator to Batch Manager. Suspend is the path the operator takes when there are still images in the batch requiring review and the operator needs to stop work (for example, at the end of a shift). The batch retains its current state and another operator can resume it from where the previous operator left off.

Complete (Ctrl+9) marks the batch complete. A batch cannot be marked Complete while images that require review are still outstanding; every image required by the workspace's review mode must be in good standing (Accepted, or Rejected with the rejection recorded for Rescan). Completing a batch triggers the QC Complete confirmation (see Completing Quality Control below) and forwards the batch to the next stage in the workflow.

Menu control:

Exit closes the right-click menu without taking action. Exit does not leave the Quality Control module.

AI Tags

AI Tags appear at the top-right of the Quality Control workspace and identify specific issues the AI detected on the current image. Each tag belongs to one of two categories, distinguished operationally by whether the issue can be fixed by rescanning the document.

Scan Issues identify problems with how the image was captured — the operator took a poor photo of the document. Scan Issues can be improved by rescanning. Common Scan Issue tags include ObjectObstruction (something blocking the document, such as a binder clip covering header text), TextCutoff (text cut off at the edge of the image due to framing), and LabelsCoveringInfo (a label covering important text).

Document Content identifies physical characteristics of the document itself that affect quality — folds, creases, stains, smudges, highlighting, faded text, and similar. Document Content issues cannot be fixed by rescanning because the underlying document has the issue. The operator's choice on a Document Content issue is to Accept the image (the issue is present but the document is the best available) or Reject the image (and seek a replacement document if one exists).

The set of available AI Tags is partly a library of common capture issues maintained by the system, and partly client-defined. Workspace administrators can create custom prompts to identify additional image quality issues specific to their projects. The same tags appear in the AI Tags section of the Properties side panel, and each tag is described in detail (with severity, location, reasoning, and recommended action) in the AI Image Quality Analysis view (Ctrl+5).

Right Panel Tiles

The right panel displays tiles for the current image. Tiles can be shown or hidden using Configure Tiles at the top right to reclaim screen space as needed.

Box Notes displays notes attached to the batch. Notes added in Review, Project Management, or previous QC sessions appear here. Operators can also add new notes from within Quality Control to record observations for downstream reviewers or the operator working the next shift.

Tags displays User tags applied to the box or its images. User tags are distinct from AI Quality Tags: AI Quality Tags are generated by the system to flag specific image quality issues (Scan Issues and Document Content), while User tags are applied by operators or administrators to categorize the box or image for project, workflow, or reporting purposes.

Omitted Images lists images that have been removed from the batch through Delete or Reject actions. The operator can restore an omitted image from this tile, which returns it to the active batch for further action.

AI Quality Score and verification

Each image carries an AI Quality Score (a percentage with a plain-language rating) and zero or more AI Tags. The score and tags are generated when the batch enters Quality Control.

The plain-language ratings (Good, Fair, and Poor) map to AI Quality Score percentage ranges configured by the client per workspace. There is no preset threshold table; each workspace defines the percentage bands that correspond to each rating. The thresholds are part of the workspace configuration under Workspace Settings. The default ratings are Good (80–100), Fair (50–79), and Poor (0–49).

Operators review the AI's flags, view each image, and either Accept (the image is acceptable as-is), Reject (the image needs to be rescanned or replaced through the Rescan Module), or apply one of the other right-click menu actions.

Workflow review modes

An administrator configures the QC workflow to require one of several review modes, set per workspace. The review mode determines which images in a batch the operator must review before the batch can be marked Complete.

100% review requires the operator to review every image in the batch. Every image must be in good standing (Accepted, or Rejected with the rejection recorded for Rescan) before the batch can be completed.

AI-tagged review requires the operator to review only the images flagged with AI Quality Tags. Images without AI tags are presumed to have passed AI scoring and do not require operator review.

Rejected-only review requires the operator to review only the images previously rejected by another user (for example, when a batch returns to QC after Rescan or from a downstream stage).

Regardless of review mode, any image the operator chooses to Reject sends the batch to the Rescan Module after the QC batch is completed. Review mode controls the set of images that must be reviewed to complete QC, not whether rejections are allowed.

Completing Quality Control

When all images required by the workspace's review mode are in good standing, the operator clicks Complete (Ctrl+9) in the right-click menu to mark the batch complete. The QC Complete confirmation appears: "You have completed QC for this record. All images have been accepted." with two options: Return to Batch Manager or Continue with next record.

Choosing Continue with next record opens the next batch in the operator's QC queue without returning to Batch Manager. Choosing Return to Batch Manager closes the completed batch and returns to the Batch Manager view.

The completed batch is forwarded to the next task in the workflow. If the batch contains any Rejected images, the next stage is the Rescan Module, where an operator replaces or rescans each Rejected image and accepts the new quality. When all Rejected images in the batch are resolved, the batch is marked Complete for Rescan and advances to the next stage in the workflow (typically Manual Redaction when AI Redaction is enabled, then Verification, then Export). If the batch contains no Rejected images, it bypasses Rescan and advances directly to the next stage.

If the operator needs to stop work before the batch can be marked Complete, the operator selects Suspend (Ctrl+8) from the right-click menu. Suspend returns the operator to Batch Manager with the batch's current state preserved. Another operator can resume the batch from where the previous operator left off.

Note: The WIB™ Unit must be connected to the internet for the Quality Control feature to work with the rework process.