Workbenches

A workbench is an interactive node that pauses a pipeline so a person can look at the data, make decisions, and (in some cases) edit it before the pipeline carries on. Where tools work automatically, workbenches are the human-in-the-loop. This section describes every workbench in Quantra in full — what it is for, every screen and panel inside it, every button, every filter, every output it can produce, and what to fill in when configuring it.

Q-Viewer

Q-Viewer is a read-only document viewer with Personally Identifiable Information (PII) highlighting. It is designed for people who need to inspect documents but who must not be able to redact, edit, or alter the originals. PII regions detected upstream are shown highlighted on the page so the viewer can see them in context.

What you'll need

• An upstream datasource that supplies the documents to view (for example, Box, Outlook, SharePoint, or a Network Drive).
• (Optional) An upstream PII Detect tool, so that PII regions are pre-marked when a viewer opens a document.

The Q-Viewer screen

Q-Viewer is a single screen with a small top toolbar and a three-panel body that you can resize by dragging the splitters between the panels.

Top toolbar

• Refresh — reloads the document list from the source.
• Close — exits the workbench.

Left panel: document list

A sortable, searchable table with five columns:

• RAG indicator — a small Red, Amber, Green (RAG) traffic-light dot summarising the document's overall health (see the colour legend below).
• Name — the filename, with a tooltip showing the full path.
• Source — an icon showing which datasource the document came from (for example, Server Message Block (SMB) network share, Box, Outlook).
• PII — a "PII" pill shown if the document contains PII.
• ROT — one or more "Redundant" or "Obsolete" pills, sortable by their score.

Above the table, a search box, a fuzzy-match toggle, and two filter checkboxes.

• Search — matches across every visible field. By default, it does substring matching.
• Fuzzy — once enabled, searches of four or more characters tolerate small typos (using edit distance), so "invoce" still finds "invoice".
• PII only — show just the documents that contain PII.
• Redundant only — show just the documents flagged as redundant.

The list paginates 200 rows at a time. Each row has a multi-select checkbox; selections are remembered as you sort, search, or filter. Clicking a row makes it the active row, opens the document in the right panel, and shows its metadata in the middle panel.

Middle panel: details

Shows every metadata field of the active row as a key/value list. Nested values (objects and lists) are rendered as expandable sections. Dates are formatted as readable timestamps and sizes are formatted in kilobytes, megabytes, or gigabytes. Internal book-keeping fields are hidden.

Right panel: document viewer

An embedded read-only viewer (the Q-DACT canvas) shows the document. Supported formats include Portable Document Format (PDF), images, videos, and text files; the right format is picked automatically. Multi-page PDFs load the first page immediately and additional pages on demand.

Above the viewer there are two toggles:

• Hide PII highlights — turns the highlighted PII regions on or off.
• Show labels — toggles the small entity-type tags that sit next to each highlighted region.

Downloads

Q-Viewer supports two download patterns:

• Single document — the active row's Download original button fetches the original file from the source.
• Bulk download — if any rows are selected, a Download selected (N) button appears. Clicking it opens a small dialog where you type a name for the archive; the files are then bundled into a single ZIP archive and downloaded.

Sign-in to cloud sources

If a document comes from a cloud source that needs Open Authorization (OAuth) sign-in (Box, Outlook, SharePoint, OneDrive, Gmail, Google Drive), Q-Viewer opens a small pop-up window for you to sign in. Once you have signed in, the workbench retries the action automatically.

Status messages

The status bar at the top shows messages such as "Loading documents...", "Loaded N documents", "Bundling N file(s)..." and "Downloaded X.XX MB", so you always know what the workbench is doing.

Keyboard shortcuts

• Esc — cancels a panel-resize drag.
• Enter — in the bulk-download filename dialog, submits the dialog.
• Esc — in the bulk-download filename dialog, closes the dialog.

Visual indicators legend

• Green dot — the document is healthy: no PII, not redundant, not obsolete.
• Amber dot — the document contains PII but is otherwise healthy.
• Red dot — the document contains PII and is redundant or obsolete; action is needed.
• Grey dot — status not yet known.
• "PII" pill — this document contains personal data.
• "Redundant" pill — the document is a near-duplicate of another in the set.
• "Obsolete" pill — the document is older than the configured retention horizon (default: six years).

Settings you provide

Setting	Required	Description
Display note	No	A label for this node on the canvas.

The other configuration — the database table that supplies the PII highlights, and which fields are shown — is set on the edge that connects an upstream datasource or database to Q-Viewer.

Outputs

• Single-document downloads — the original file, fetched live from the source datasource.
• ZIP archive — a single archive of every document the user has selected, with a name the user picks.

Common uses

• Letting auditors or external reviewers inspect a document set without giving them the ability to redact or edit.
• Spot-checking PII detection quality across a backlog of documents.
• Producing a controlled bulk download of a chosen set of source documents.

Q-ROT

Q-ROT is the workbench for Redundant, Obsolete, Trivial (ROT) data analysis. It looks across a document collection and classifies every item using a Red, Amber, Green (RAG) traffic-light scheme — Red items are clearly redundant, obsolete, or trivial, Amber items are borderline, and Green items are healthy. Q-ROT is most often used as part of General Data Protection Regulation (GDPR) data-minimisation work.

What you'll need

• An upstream datasource that supplies the documents to analyse.
• (Optional) An upstream Hash tool so that near-duplicates can be detected.

The Q-ROT screens

Q-ROT has four tabs — Dashboard, Documents, Settings, and Help.

Dashboard tab

An at-a-glance overview made up of eight summary widgets:

• Total Files — how many files have been analysed.
• Red count — files that need action, shown in red.
• Amber count — files that need a closer look, shown in amber.
• Green count — files that are healthy, shown in green.
• Redundant — how many files are flagged as duplicates or near-duplicates.
• Obsolete — how many files are older than the retention horizon.
• Trivial — how many files are flagged as trivial system files.
• Potential Savings — the total disk space, in kilobytes, megabytes, or gigabytes, that could be reclaimed by removing the Red files.

Below the widgets is the RAG distribution bar, a stacked bar that shows the share of Red, Amber, and Green files visually, with a legend and percentages.

Run analysis

The Dashboard's status section shows the timestamp of the most recent analysis (in green if one has run, in amber if it has not). Two action buttons sit next to it:

• Run Analysis — opens a dialog with two choices: Redo (re-analyse every file from scratch) or Update (analyse only the files that have arrived since the last run). The dialog has a Cancel button.
• Export Report — downloads a Comma-Separated Values (CSV) report of every analysed file with its RAG status, ROT flags, and key metadata. The downloaded file is named rot_analysis_report.csv.

Documents tab

A three-panel analysis view modelled on Q-Viewer, with a list of files on the left, a metadata pane in the middle, and a Q-DACT preview on the right. The list shows each file's RAG dot, a category badge (Redundant, Obsolete, Trivial, or Clean), and key metadata.

Above the list:

• Search box — full-text search across every visible field.
• Filter — opens a dialog where you can build more specific filters (for example, "show only Red files" or "files older than 5 years").

Other table features include sortable columns, a sticky header, multi-select checkboxes, and row-hover highlighting. The detail panel renders nested fields as expandable sections.

Settings tab

A summary of the configuration the canvas-time settings have set: thresholds, default retention, trivial patterns, and so on.

Help tab

Built-in usage guidance describing how to interpret RAG categories and how to use the analysis dialog.

Status messages and progress

Q-ROT shows a loading spinner overlay while the analysis is running, and a status bar with colour-coded messages (red for errors, green for success, neutral for information). The "Last analysis" timestamp turns green once an analysis has completed.

Visual indicators legend

• Red dot / Red badge — action required: the file is clearly redundant, obsolete, or trivial.
• Amber dot / Amber badge — review needed: the file is borderline.
• Green dot / Green badge — healthy.
• Grey dot — not yet analysed.
• Redundant / Obsolete / Trivial pills — the specific reason a file has its colour.

Settings you provide

Setting	Required	Description
Near-duplicate threshold	No	How aggressively to call two files near-duplicates. Lower numbers are stricter; higher numbers are looser. Defaults to 30 (strict). 60 is moderate; 100 is loose.
Default retention (days)	No	How long, by default, a document is allowed to be kept. Anything older than this is treated as obsolete. Defaults to 2,555 days (about seven years).
Grace period (days)	No	Documents that will become obsolete within this window are flagged Amber instead of Red. Defaults to 90 days.
Trivial size threshold (bytes)	No	Files at or below this size are flagged as trivial. Defaults to 1,024 bytes.
Trivial file extensions	No	A comma-separated list of file extensions to treat as trivial. Defaults to .tmp,.bak,.log.
Trivial name patterns	No	A comma-separated list of filename patterns to treat as trivial. Defaults to ~$,thumbs.db,.ds_store,desktop.ini.
Exclude file types	No	A comma-separated list of file types to leave out of the analysis altogether.

Outputs

• Dashboard counts and the RAG distribution bar — the live, on-screen summary of the analysis.
• RAG classification per document — flags written back to the document set so downstream tools and workbenches can act on them.
• Cluster groupings — near-duplicate documents grouped together for review.
• Analysis report — a downloadable CSV (rot_analysis_report.csv) of every file with its RAG status, flags, and metadata.

Common uses

• Identifying out-of-date and duplicate content on a shared drive or document library.
• Producing a defensible record of what was deleted as part of a data minimisation exercise.
• Surfacing trivial system files so they can be excluded from downstream processing.

Q-SAR

Q-SAR is the workbench for fulfilling a Subject Access Request (SAR) under General Data Protection Regulation (GDPR) and similar regimes. It presents the documents and structured records gathered for a SAR in a searchable, filterable, three-panel viewer, and supports the row-by-row selection that downstream redaction and release tools depend on.

What you'll need

• One or more upstream datasources or tools that have gathered the documents and records that fall in scope of the request.

The Q-SAR screens

Q-SAR has three tabs.

Overview tab

The starting page. Shows the workflow status, any guidance the project has set, and a button to switch to the Viewer tab.

Viewer tab (standard layout)

The main working screen. The header shows the title, an indicator of the current layout ("Standard layout"), and a button to switch to the Explorer view.

Below the header, a Data info section shows:

• The filename or table name being shown.
• Row and column statistics, for example "523 rows × 12 columns (database)".
• Refresh Data — reloads from the source.
• Export Filtered — downloads the currently filtered records as a CSV file (initially hidden until a filter is applied).
• Column Actions — a small toolbar with Auto-Fit All Columns, Reset Column Widths, and the hint "Drag column borders to resize, double-click to auto-fit".

Next, the Search and Filter bar:

• Column dropdown — ALL (search every column) or pick a specific column.
• Operator dropdown — with twelve operators: contains, equals, starts with, ends with, regex, greater than, less than, greater than or equal, less than or equal, between, not empty, and is empty.
• Value input — appears or hides depending on the operator, with hints for ranges and regular expressions.
• + Add Filter — stacks another filter on the current set.
• Clear All — removes every active filter.
• Fullscreen — expands the workbench to fill the screen.
• Active Filters — each active filter is shown as a tag with an "×" to remove it individually.
• Status indicator — "Showing M of N rows (X%)" so you can see at a glance how aggressive your filters are.

The body is a three-panel layout with draggable splitters between the panels:

• Records (left) — a sortable table of items with a count badge of how many match the current filter. The footer of this panel has the record count and an Export button. The table paginates 200 rows at a time and supports multi-select checkboxes; the active row is highlighted.
• Details (middle) — key/value metadata for the active row in a multi-column layout, with the empty state "Select a row to view details".
• Preview (right) — a document preview (PDF, image, text, video) with the filename in its header and the empty state "Select a row to preview".

Viewer Explorer tab (advanced layout)

An alternative layout of the same data for users who prefer a more table-centric view. The administrator can configure the workbench to default to either layout; the two are otherwise interchangeable.

Sign-in to cloud sources

If a record points to a document that lives in a cloud source needing OAuth sign-in, Q-SAR opens the same kind of pop-up window as Q-Viewer.

Keyboard shortcuts

• Ctrl + F (or ⌘ + F on Mac) — focuses the search value input.
• Esc — cancels a panel-resize drag.

Settings you provide

Setting	Required	Description
Display note	No	A label for this node on the canvas.
Table columns	No	Which fields to show as columns in the main table, separated by commas. Use * to show every available field. Defaults to *.
Detail columns	No	Which fields to show in the metadata panel when a row is selected, separated by commas. Use * to show every field. Defaults to *.

Outputs

• Filtered CSV export — the currently filtered rows are exported as filtered_data_<timestamp>.csv.
• Selected rows — the rows ticked in the table are made available to downstream tools (such as PII Detect, Review, and SAR Release) as the in-scope set for the request.

Common uses

• Reviewing the documents and records found in response to a Subject Access Request before they are redacted and released.
• Letting a Data Protection Officer (DPO) sign off the in-scope items for a request.
• Producing the filtered CSV that downstream redaction and release tools act on.

Q-RoPA

Q-RoPA is the workbench for maintaining the Record of Processing Activities (RoPA) that GDPR Article 30 requires every controller and processor to keep. Unlike the other workbenches, Q-RoPA is not primarily a document review interface — it is a structured register of how an organisation processes personal data, who it shares it with, on what legal basis, and for how long.

What you'll need

• One or more upstream datasources or tools that supply the documents and processing metadata Q-RoPA will use to populate the register.

The Q-RoPA screens

Q-RoPA has six tabs — Dashboard, Register, Documents, Clusters, Analysis, and Admin — plus a Help tab.

Dashboard tab

Eight summary widgets, each colour-coded so risk areas stand out:

• Activity Count — how many processing activities the register currently holds.
• Document Count — how many source documents have been ingested.
• PII Detections — how many pieces of PII have been found across all documents (amber if greater than zero).
• Cluster Count — how many groupings of related documents have been created.
• Documents Pending Analysis — documents waiting to be processed (red if greater than zero).
• Unclustered Documents — documents that have not yet been grouped (red if greater than zero).
• Activities Missing Retention — processing activities without a retention period set (red if greater than zero).
• Special Category PII — how many special-category items (for example, health data, biometrics, ethnicity) have been found (red if greater than zero).

The dashboard sidebar offers two analysis actions:

• Ingest — pulls in any new documents from upstream and runs PII detection on them.
• Cluster — groups related documents into processing activities using the AI helper.

Register tab

The heart of Q-RoPA: a table of every processing activity. The columns are:

• ID — a short numeric identifier.
• Activity Name — the name of the processing activity.
• Department — the team responsible.
• Status — a coloured badge: Draft (grey), Active (green), Under Review (amber), or Archived (red).
• Legal Basis — one of the six GDPR legal bases: Consent, Contract, Legal obligation, Vital interests, Public task, or Legitimate interests.
• Risk Level — a coloured badge: Low (green), Medium (amber), High (orange), or Critical (red).
• Confidence — how sure the platform is about the activity, with a coloured dot: low (yellow), medium (blue), or high (green).

Click any column header to sort by it (click a second time to reverse). Click any row to expand the inline edit form, which exposes every Article 30 field for that activity:

• Activity Name, Department, Description / Purpose, Status, Legal Basis
• Controller Name, Controller Contact, DPO Name, DPO Contact
• Categories of Data Subjects, Categories of Personal Data, Categories of Recipients
• International Transfers, Safeguards
• Retention Period, Risk Level, Technical and Organisational Measures, Confidence, Data Protection Impact Assessment (DPIA) Required (Yes / No / Unknown), Notes

Three buttons sit at the bottom of the form:

• Save — commits changes to the register.
• Cancel — discards the unsaved edit.
• Delete — removes the activity from the register, with a confirmation prompt.

The Register toolbar also has:

• New Activity — opens a blank edit form for a new processing activity.
• AI Review — opens a modal that lists each activity together with the AI helper's findings, suggested risk level, and suggested improvements.
• Export PDF — downloads the register as a PDF report (ropa_register.pdf).
• Export CSV — downloads the register as a CSV file (ropa_register.csv).

Documents tab

A table of every source document Q-RoPA has ingested, with the link to the processing activity it was assigned to and a summary of the PII detected in it. Rows can be selected to drive cluster operations.

Clusters tab

The groupings the AI helper has built. Each cluster shows the documents inside it, a similarity score, and three operations:

• Merge — combine two clusters into one.
• Split — break a cluster into multiple clusters.
• Move — move a document from one cluster to another.

Analysis tab

A history of analysis runs, with a progress bar for any run currently in flight. Each entry shows when the run started, how long it took, how many items it processed, and a status badge: in progress, success, failed, or partial.

Admin tab

Visible only to administrators and DPOs. Contains threshold settings, the legal-basis taxonomy, retention policy defaults, and the export configuration for the register report.

Help tab

Built-in usage guidance for the workbench.

Visual indicators legend

• Status badge — Draft (grey), Active (green), Under Review (amber), Archived (red).
• Risk badge — Low (green), Medium (amber), High (orange), Critical (red).
• Confidence dot — low (yellow), medium (blue), high (green).
• Special-category PII badge — flags health, biometric, ethnicity, religion, and other special-category data.
• "Missing retention" warning — an activity has no retention period set.

Settings you provide

Setting	Required	Description
Display note	No	A label for this node on the canvas.
Organisation identifier	Yes	A short identifier for your organisation, for example org-acme-corp. Every project that uses the same identifier (and the same database) sees and contributes to the same register.
Organisation description	No	A short description of what the organisation does, for example "Mid-size insurance broker. Processes policyholder PII, claims, and correspondence." This description is used by the AI clustering to produce more accurate processing activities.

Outputs

• The processing activities register — the live, persistent record of every processing activity, viewable on screen and editable inline.
• PDF register report — a printable Article 30 report (ropa_register.pdf) suitable for submission or audit.
• CSV register export — the same data as a CSV file (ropa_register.csv) for use in other tools.
• Cluster decisions — merge, split, and move actions are written back to the register so subsequent runs build on them.
• AI Review summary — the on-screen list of findings, risk levels, and suggestions for each activity.

Common uses

• Maintaining a GDPR Article 30 Record of Processing Activities for the whole organisation.
• Supporting a Data Protection Officer in oversight, sign-off, and reporting duties.
• Producing a defensible record of why each kind of personal data is processed.
• Generating audit-ready PDF and CSV exports of the register.

Q-NHS-SAR

Q-NHS-SAR is the National Health Service (NHS) variant of Q-SAR. It works in the same three-panel way as Q-SAR — a table of items, a metadata panel, and a document preview — but with column defaults, metadata field mappings, and PII rules adjusted for healthcare. Use Q-NHS-SAR when the request is for clinical and patient-administrative records rather than general business records.

What you'll need

• One or more upstream datasources or tools that have gathered the patient records and supporting documents in scope of the request.

The Q-NHS-SAR screens

Q-NHS-SAR has the same three tabs as Q-SAR — Overview, Viewer (standard layout), and Viewer Explorer (advanced layout) — plus a Help tab.

Overview tab

The starting page, with workflow status and any guidance the project has set.

Viewer tab

The same layout as Q-SAR — data info section, search and filter bar with the same twelve operators (contains, equals, starts with, ends with, regex, greater than, less than, greater than or equal, less than or equal, between, not empty, is empty), three-panel body (Records, Details, Preview), and column actions (Auto-Fit, Reset, drag-resize) — but with NHS-specific defaults.

What is different from Q-SAR

• Default columns — NHS Number, Patient Identifier, Facility or NHS Trust, Record Date, and Document Type are visible by default, in addition to the standard columns. The administrator can override these via the canvas-time settings.
• Default detail fields — the metadata panel shows NHS-specific fields such as NHS Number, Medical Record Number (MRN), Episode reference, and Date of encounter, alongside the standard fields.
• Identifier badges — rows that have been validated as carrying a real NHS Number or MRN show a small badge so reviewers can confirm they have the right patient.
• PII rules — the upstream PII detection follows NHS information-governance guidance rather than generic GDPR rules, so identifiers such as NHS Numbers and MRN values are explicitly recognised.

Viewer Explorer tab

The same alternative table-centric layout as Q-SAR, with NHS column defaults applied.

Sign-in to cloud sources

If a record points to a document that lives in a cloud source needing OAuth sign-in (for example, an NHS-connected document library), Q-NHS-SAR opens a pop-up window in the same way as Q-Viewer and Q-SAR.

Keyboard shortcuts

• Ctrl + F (or ⌘ + F on Mac) — focuses the search value input.
• Esc — cancels a panel-resize drag.

Settings you provide

Setting	Required	Description
Display note	No	A label for this node on the canvas.
Table columns	No	Which fields to show as columns, separated by commas. Use * to show every field. Defaults to the NHS-specific column set.
Detail columns	No	Which fields to show in the metadata panel when a row is selected, separated by commas. Use * to show every field. Defaults to the NHS-specific field set.

Outputs

• Filtered CSV export — the currently filtered rows, with NHS-specific columns, exported as filtered_data_<timestamp>.csv.
• Selected rows — the rows ticked in the table are made available to downstream tools (PII Detect, Review, SAR Release) as the in-scope set for the request.

Common uses

• Fulfilling patient access requests under NHS information-governance rules.
• Reviewing clinical and administrative records before redaction and release.
• Producing the filtered CSV that downstream NHS redaction and release tools act on.