Punchcard product updates

New

Drop files anywhere on the Workpaper — they sort themselves

Drag files onto any row in the Workpaper table to upload and attach them to that specific selection. Drop files anywhere else on the page and each file is automatically sorted into the best-matching document group — no manual assignment needed. A new "Unsorted" column on the right side of the table tracks files as they process, with a per-file status popover showing progress and an option to remove files that couldn't be matched. The reference-file panel now only expands when you deliberately hover the paperclip icon, so dropping files anywhere else on the page always goes to the auto-sort flow.

CoAudit can answer questions about your audit data

The CoAudit chat assistant now has direct access to your audit portfolio. Ask "show me all overdue tasks," "what's the oldest outstanding client request?" or "find tasks assigned to me in this engagement" — CoAudit queries your clients, engagements, tasks, activity log, and workpaper comments to answer. Available in both standalone chat and task-embedded chat.

Financial Statement Review: footing and recalculation in one unified procedure

Footing (sum checks) and Recalculation (multiplication and division checks) are now presented as a single "Recalculation" procedure in the sidebar and cover sheet. Draw your operands as usual — the system detects the operation type automatically and routes the result to the correct check. Your firm's existing footing glyph applies to all results. The Manage Tick Marks dialog no longer shows the now-redundant duplicate entries from before the merge.

Financial Statement Review: dedicated "Add a comment" button

A new message icon in the PDF toolbar puts you directly into comment-placement mode — click any spot on the document to start a thread, without navigating through the tick-mark picker. Previously, adding a standalone document comment required going through the tick-mark flow, and in math-validation mode it was not possible at all. Standalone comments are now visible and accessible across every sidebar mode.

---

Improvements

• Engagement Settings empty states are now clickable — "Add your first category" and "Add your first variable" now open the creation form when clicked. Previously they looked like buttons but did nothing.
• Activity stream file names open the viewer — clicking a file name in a task's activity feed now opens that file in the document viewer. Deleted files remain as plain text.
• Activity feed updates after file changes — uploading or deleting a reference file now refreshes the activity stream immediately, without requiring a full page reload.
• Archived engagements display correctly — engagements in Archived status now show as "Closed" across the client list, engagement table, and sidebar, rather than appearing as "Planning." Archived is also now a selectable option in the engagement status picker.
• Financial Statement Review outputs in the document's language — tick-mark notes, group labels, and analysis text now match the language of the financial statement being reviewed. A regression had been causing English-document output to appear in Spanish or German for many files across multiple firms.
• FSR accuracy improvements for German and EU-format statements — Financial Statement Review now correctly skips index pages, asset registers, and other appendix tables rather than flagging their entries as financial exceptions. Single-column numeric tables in EU decimal format are now correctly extracted and validated.
• FSR navigation after confirming a result — after confirming a footing or recalculation result, the sidebar now correctly advances to the next item in sequence, including sibling relationships on the same cell (such as both a row sum and a column crossfoot on one number).
• PDF viewer controls always respond — the zoom, page navigation, and fullscreen controls now respond to clicks even when the toolbar has faded from inactivity. Hovering them brings the toolbar back.
• @mention comments always save — if a notification email to a mentioned colleague can't be delivered, the comment saves successfully regardless. Previously a delivery failure would cause the whole comment to appear to fail.
• "Paste from clipboard" shows a loading indicator — the paste button in the sampling table now shows a spinner while waiting for clipboard permission, so it no longer looks unresponsive during the browser prompt.
• Citations in AI responses render correctly — inline citations in CoAudit responses now display properly even when some references can't be resolved, and unresolved placeholders no longer appear as stray characters in the message text.

---

Bug Fixes

• Fixed a hang that could occur when selecting a match on certain workpaper files — the action now responds quickly rather than waiting several minutes before failing.
• Fixed an issue where navigation could freeze for users with Cyrillic, Asian-script, or emoji characters in engagement or task names.
• Fixed large documents occasionally failing to complete text recognition when multiple documents were being processed at the same time.

Office Add-ins — Redesigned Review Panel

• Citation details now appear in a docked bottom panel below the document viewer, so the source document and your review notes are visible at the same time. Toggle the panel open or closed with a button in the pane controls, drag the divider to adjust its height, and the add-in remembers your preference across sessions. The panel works the same way in both Excel and Word.
• Validate and Flag are now available directly in the panel — no need to reach for the ribbon for single-citation verdicts. Unreviewed citations show Validate and Flag buttons in the bottom panel; once a verdict is set, the button becomes a status badge (Validated / Flagged) with a one-click clear option. The ribbon Review group still handles bulk selections in Excel.
• New Settings button (gear icon) in the ribbon lets you re-link a bound workbook or Word document to a different engagement or task without closing the pane. Use the familiar cascading picker to select or create a client, engagement, or task.
• Sign-in is more reliable. If Microsoft's authentication flow takes an unusual path, the add-in now detects the completed sign-in and returns to the pane automatically rather than appearing stuck. If the pane ever lands on the wrong page, clicking any ribbon button brings it back immediately.

Workpaper

• Clicking a cell now opens the correct source document. The document viewer opens to the file where the cell's primary citation is highlighted — not to a fallback file that may have nothing to do with that cell's evidence. Citations that reference a file from a different part of the engagement also load correctly.
• The collapse/expand arrow on datasource column groups now points in the correct direction after each click.

Navigation

• Browse sibling engagements from the breadcrumb — hovering the engagement name in the breadcrumb opens a panel listing all engagements under the same client, sorted by most recently updated. Each row shows the engagement's status, title, and a color-coded bar of task and request counts by status. Clicking any row navigates to that engagement while preserving your current Tasks or Requests view.
• Refreshed breadcrumb design — the breadcrumb adopts a new chip style with cleaner idle, active, and hover states. On the engagement board, the audience label ("Tasks" or "Requests") no longer appears as a trailing segment; the trail ends at the engagement name.
• Board Display menu no longer freezes the page — the column, row, and ordering settings panel now stays open while you select options from its dropdowns, and the board behind it remains scrollable. Previously, opening the Display menu made the entire page unscrollable until it was closed.

CoAudit AI

• Approval cards no longer end the AI session — when CoAudit reaches a step requiring your approval, it pauses inline and resumes automatically once you approve or decline. You no longer need to type anything to continue the AI's work after acting on an approval.
• Decline reason improves the AI's next attempt — when you decline an approval, you can optionally explain why. The AI receives your explanation and adjusts its next proposal, eliminating the need for a separate follow-up message.
• Excel add-in: approval cards stay visible across all tabs — a pending approval now appears above the chat, Documents, and Viewer tabs so you won't miss it if you switch views while the AI is waiting.
• Excel add-in: workbook operations run without interruption — the AI can now read and write workbook cells across a long authoring session without pausing between operations, making complex workbook tasks faster to complete.

Firm Library

• Contextual empty states — when a tag filter returns no results, the library shows "No documents match the selected tags." When a text search returns nothing, it shows "No matching documents found." The upload prompt appears only when the library is genuinely empty with no active filters, so filtered views no longer suggest your library is empty.
• Preview files uploaded by colleagues — firm members can now open and preview documents that a colleague uploaded to the shared library. Previously, opening another person's file showed a "Failed to load file preview" error.

File Uploads

• Bulk uploads no longer silently drop files — uploading a large batch of files at once no longer fails partway through and discards some files. A real production session had previously lost files mid-batch; this is resolved.

Workpaper Automation

• AI processing recovers from concurrent edits — if workpaper records are deleted while the AI is actively processing them, the remaining extracted results are now preserved rather than the entire batch being discarded.

Accessibility

• Screen reader labels for draw and comment controls — the comment toggle, draw line/arrow button, and split-button in the document viewer toolbar now announce their names and active states correctly to screen readers.

Excel Add-in — Workbook Authoring

• Writing to a workpaper cell now requires a single approval, not three or more. Previously, filling one cell correctly could require separate approvals for the value, the formatting, and the citation. All of these are now batched into one step, so you approve once and everything is applied together.
• AI can read a cell's current contents — including its formula, citation, and formatting — and update only the parts that need to change, leaving everything else untouched. This makes targeted corrections predictable and avoids accidentally clearing a reviewer's existing work.
• Cells you've already filled are protected from silent overwrites. If AI tries to write over a cell that contains your work, it stops, shows you what is already there, and waits for you to confirm before replacing anything. When you do approve a replacement, the confirmation card leads with a clear notice that existing data will be replaced.
• AI can now attach multiple evidence citations to a single cell. When a figure rests on more than one piece of supporting evidence — for example, a ship date and freight terms from two different lines of the same document — each source gets its own clickable highlight in the document viewer.
• When AI cites a passage it cannot locate in the source file, it now reports exactly which quote didn't match and suggests how to retry, rather than silently appearing to succeed with no highlight to click.
• The approval mode selector is now labeled Manual and Skip all approvals (previously "Ask" and "Approve All"), with a raised-hand icon for Manual and a warning-triangle for Skip, making the difference clear at a glance. The collapsed toggle shows the short form (Manual / Skip) to fit the task pane.
• File imports always require their own approval, even when a write session is active. Other workbook writes — values, formatting, citations — still proceed without interruption once a session is approved; importing a document group always shows a separate approval card.

Excel Add-in — Document Manager

• Two tabs now appear over the file list: Cited and Text. The Cited tab shows which files have evidence linked to workbook cells. The new Text tab groups files by text-recognition status — successfully processed, still recognizing, failed, and unsupported — with a red badge for failures and a purple badge for files still being processed, visible from the Cited tab so extraction problems are never silent.
• Text recognition status refreshes automatically while files are processing: every two seconds for the first 30 seconds, then every five seconds up to two minutes, then every 15 seconds. A "Check again" link appears if recognition hasn't finished within that window.
• File names appear in the Document Manager as soon as you select them for upload, with a spinner in place of the file-type icon, so you can see what is in flight without waiting for the transfer to complete.
• Drag files, folders, or ZIP archives directly onto the Document Manager to upload them. Dropping onto an open document group uploads into that group; dropping above the list opens the standard destination chooser.
• The region selector (US / EU) now appears only when needed — during sign-in or when the pane cannot connect — rather than displaying a banner on every launch.

CoAudit Citations

• Multi-word quotes now match correctly in source documents. Citing a phrase such as "Pay Date: 12/22/2025" previously failed to find a match because the quote spans multiple words in the document's underlying text, producing a full-page highlight with nothing to click. The highlight now lands precisely on the quoted text.
• Citation highlights cover the full quoted phrase, not just the first word. A three-word quote now produces a single highlight spanning all three words.
• False citation matches are prevented. A quote can no longer match text where it appears only as part of a longer phrase — for example, "Pay Date" will not highlight "Repay Dated."

Karbon Integration

• Firm admins can now connect their Karbon integration without contacting support. Previously this required a backend command run by an operator. Firm admins can now go to Admin → Integrations → Karbon and paste their Karbon access key directly. The key is verified against Karbon before being saved, so a typo cannot replace a working connection.

Financial Statement Review

Merge tables that span page breaks

• Select two adjacent-page tables in the review sidebar and choose a merge strategy: Vertical (the table continues across the break) or Horizontal (columns are split side by side).
• A live preview shows row accounting, a list of any values that would not carry over (named by amount, not just counted), and whether column headers can be reconciled — all before you commit.
• Merges are non-destructive: existing tick marks, variance analyses, and comments are preserved and carried to the merged result.
• A dashed seam line in the PDF marks where tables were joined. Undo a merge by hovering the seam and clicking Unmerge, or using the revert control in the sidebar list. Math validation re-runs automatically after each merge or unmerge.

Correct a misread presentation scale

• Every Foot/Crossfoot calculator now shows a unit chip (e.g., "TEUR", "thousands") auto-detected from the table header.
• Click the chip to correct a wrong detection — for example, changing "full amounts" to "thousands" when the statement is in TEUR. All findings in that calculation re-score immediately, and tick marks on findings whose verdict changed are flagged for recheck.
• A filled chip indicates a system-detected unit; an outlined chip means you set it manually. A "Reset to detected" option restores auto-detection.

PDF text search

• Press Cmd/Ctrl+F while viewing a PDF to open a find bar with previous/next navigation through all matches.
• In dual-pane Current Year / Prior Year view, the shortcut routes to the pane you last interacted with.
• The shortcut is not intercepted when the cursor is inside a text field such as a comment box.

AI analysis now matches the document's language

• Internal Consistency grouping labels, Foot/Crossfoot correction reasoning, and Prior Year comparison analysis are now written in the document's own language — German, Spanish, Italian, and others — rather than always in English.
• Table titles extracted during enrichment are preserved verbatim rather than translated.

Accuracy improvements

• Fewer false math exceptions: the math correction step now discards proposed total-row compositions that fail to reconcile on any constraint before flagging them. This removes a class of false exception marks on correct statements, such as the example where two unrelated line items were incorrectly marked as totals.
• Percent-change columns with negative bases: whether a "% change" figure uses a signed or absolute-value convention is now inferred from the document by majority vote rather than assumed, removing false exceptions on correctly-stated rows.
• Multi-line row labels in European-format statements — including labels that open with a currency-prefixed value like "−€1,234.56" or "€−1,234" — are now correctly joined into a single row, preventing false footing exceptions on wrapped captions.

---

CoAudit Chat in Excel and Word

Workbook authoring (Excel)

• CoAudit can now read and write the live Excel workbook — cell values, formulas, number formats, fill colors, merged cells, and citations — directly from the task pane.
• Ask CoAudit to build a cited workpaper schedule: it reads source PDFs, extracts figures with page references, writes values and formulas into cells, and attaches citations so each number traces back to its source page in the PDF.
• Write Sessions: approve a described block of work once ("populate the detail sheet for 26 paystubs, with citations") and all workbook writes in that scope proceed without interruption. Sharing a document group from another task still asks individually.
• Approve All: a toggle beside the Citations control grants a standing approval for the current conversation, so nothing interrupts a long unattended run. Resets when the tab is closed.

CoAudit chat now in Word

• The Word task pane now includes CoAudit chat with the same AI assistant as the web app. Citation review and source evidence are unchanged; workbook authoring tools are Excel-only.

Document group management in the pane

• Create, rename, and delete document groups from the task pane without opening the web app. A 3-dot menu on each group offers Rename and Delete, with a file count shown in the confirm dialog before deleting.
• The + tile offers two options in one click: create a new group or import an existing group from another task in the engagement.
• Empty document groups now appear immediately after creation rather than staying invisible until a file is uploaded into them.
• Files uploaded through the pane now trigger text recognition automatically, matching web upload behavior.
• Renaming or removing a shared document group from an importing task no longer affects the source task or any other task that shares the group.

Other add-in improvements

• The ribbon button that opens the citation and evidence surface is now labeled "Viewer" (previously "Review") to distinguish it from the Validate / Flag / Clear review actions.
• The import panel navigation has been simplified: the engagement name and a back chevron appear in the header rather than in the scrolling body, so the context label stays visible as you drill into folders.

---

Workpaper

• PDF text search: press Cmd/Ctrl+F while a PDF is open in the workpaper to search the document's recognized text with previous/next navigation.
• Removing the last client-uploadable document group now presents an explicit choice: move the task to the internal Tasks board, or keep it as a client request with no upload destination. The outcome is no longer decided automatically without input.

---

Reliability

• Certain files were permanently undownloadable if their name contained a space near a dot (e.g., "Report pg. 2"). These files now download correctly.
• Navigating directly to an engagement board URL could occasionally leave the page blank if a navigation attempt was dropped during load. The page now shows a spinner and retries until navigation succeeds.
• MUS formula panel: the formula result shown in the Monetary Sample Size panel now matches the formula's own arithmetic rather than the final selected row count, which could differ on small populations.
• File moves within the workpaper now complete atomically, so a move on a large file with many records no longer risks leaving data in a half-migrated state if it runs long.

Office Add-ins (Excel & Word)

• Word ribbon now matches Excel — The Word add-in ribbon now includes the same controls as Excel: CoAudit (AI chat), Sum (draw boxes across pages to total values), and Clear citations. These were previously absent from the Word ribbon, requiring workarounds.
• CoAudit / Review tabs in full-screen chat — When the AI chat panel takes the full screen on a narrow pane, tabs at the top let you switch between CoAudit and Review without navigating away.
• Cross-document search cycling — After searching across your source documents, the next/previous arrows now cycle through results from all files in sequence, opening each document at the matching location as you step through.
• Sum mode keeps drawn boxes visible — While drawing additional sum components, citation boxes already saved for that cell stay highlighted in the viewer so you can see exactly what you have already captured before adding more.
• Upload destination redesigned — Choosing where to upload files now shows a card list with an "Upload Here" button for each folder, matching the look and feel of the document import flow.

PDF Viewer

• Comments on drawn lines and arrows — Selecting a drawn line or arrow in the PDF viewer now opens the same action menu as a citation number, including the ability to add and view comment threads. Previously, drawn marks only offered a delete option.
• Sum calculator matches cell currency format — The running total in the sum calculator now uses the same number and currency format as the workpaper cell it feeds, so the displayed values are consistent.

Workpaper & Automation

• Automation conflicts give a clear retry message — When a match or automation run encounters a conflicting in-progress operation (such as a file deletion), it now fails within seconds with a clear message to retry, rather than hanging silently for several minutes before erroring.
• Manual footings are fully editable — Manually added footings in the math sidebar now support the same edits as pipeline-discovered footings: you can correct the stated result and adjust operator signs, and the verdict updates automatically when the numbers change.

Reliability

• Bulk uploads recover from brief network interruptions — Large file uploads now automatically retry when a transient network issue occurs, instead of failing and requiring a complete re-upload.
• Files that finish processing late no longer stay stuck — OCR occasionally completed after the initial wait window, leaving files in a permanent "timed out" state. These files now automatically show as ready when processing finishes.
• Deleting tasks with many files now succeeds reliably — Tasks with a large number of attached files now delete completely rather than timing out partway through.

Financial Statement Review

• More consistent table structure reading — The AI that identifies rows, headers, and columns in financial statement tables has been upgraded to a significantly more reliable approach, reducing misclassified rows and inconsistent results across successive runs on the same document.
• Improved accuracy on German financial schedules — Several patterns that caused false exceptions on German-format statements are now handled correctly, including schedules where the total column appears to the left of the detail columns, and expense or provision line items that were incorrectly flagged as subtotals.
• Footing discovery improvements — The system now correctly proposes footings for single-member section subtotals, balance-sheet identity rows (where two total rows are cell-for-cell equal), and paragraphs containing label-less totals that were previously invisible to the discovery engine.

Workpaper

• Export menu labels now match what you receive: the "Workpaper Only" and "Files + Workpaper" export options now show the correct file format — .zip when sampling reports are bundled in, .xlsx otherwise. The label and the download now always agree.

CoAudit Chat

• Feedback is now confirmed with a message: clicking thumbs up or thumbs down on a CoAudit response now shows a brief confirmation ("Thanks for your feedback") so you know your rating was recorded.
• The AI recovers from spreadsheet column name mismatches: when querying or filtering spreadsheet data, the AI now finds columns even when capitalization or spacing differs from the file's headers, and tells you the available column names when it still can't resolve one — so fewer repeated failed attempts before it finds the right data.
• Documents with spaces or special characters in their names now work correctly: files whose names include spaces, parentheses, or similar characters no longer cause issues when the AI processes or delivers them.
• Engagement variables are resolved more reliably in task chats: the AI can now look up engagement-wide variables correctly across all task types, including scenarios that previously returned an error.

Financial Statement Review

• Citation highlights now point to the correct text: an issue where a citation sometimes highlighted text in the wrong part of a page — such as a character inside a row label instead of the actual value — has been fixed.
• Fewer false exception flags on valid line items: several categories of incorrectly flagged amounts are no longer surfaced as exceptions, including roll-forward opening balances, ratio and rate rows, amounts in limit or percentage columns, and certain narrative-captioned items.
• Progress tracking during analysis is accurate: the status indicator shown while Financial Statement Review is running now reflects each step's real progress. Previously, the indicator could show outdated information during a re-run or after reconnecting.
• Analysis completes reliably after a re-run: a case where the progress spinner would hang indefinitely following a failed re-run has been fixed.

Platform

• Invite dialog shows the correct client name: the dialog for inviting team members or clients in Engagement Settings now reliably displays the correct client organization name. It previously showed "No client" in some situations.
• Library content search closes correctly when dismissed: clearing or closing a content search while results were still loading now dismisses the panel as expected instead of leaving it stuck open.
• Login restored for users with duplicate account records: a small number of users who were unable to access Punchcard due to a duplicate account record can now log in successfully.

Financial Statement Review

Group-level decisions in Internal Consistency and Prior Year Consistency

• Settling a group is now a single action. Confirming or flagging a number group resolves all its occurrences — current-year and prior-year — simultaneously. Previously each occurrence had to be handled individually, and prior-year numbers had no controls at all, leaving some groups stuck showing a partial verdict indefinitely.
• Groups display the number of occurrences they cover rather than a running reviewed/total count. Sidebar progress now tracks groups, not individual occurrences — one group is one decision.
• Each occurrence can be removed from its group independently. Every row, including prior-year rows, now has a "Remove from group" action to dispute a number's membership without changing the group's verdict for the remaining occurrences.

Recalculation: multiplication and division relationships

• Financial Statement Review now automatically detects when a value is mathematically derived from other cells via multiplication or division — for example, a margin percentage calculated from revenue and expenses, or a contributions figure expressed as a percentage of covered payroll. Tick marks are suggested when the relationship holds and flagged when the calculation does not reconcile.
• Only well-supported relationships are surfaced; proposals without enough corroborating evidence are filtered out automatically, keeping exception marks meaningful.

Number format detection per statement

• FSR now detects whether each financial statement uses US or European number formatting from the document itself, rather than always applying the firm's default. This prevents incorrect footing results when a firm audits statements prepared in a different country's format — for example, a European firm reviewing a US-formatted balance sheet.

Fewer false exceptions in Internal Consistency

• IC no longer raises exceptions on numbers that are part of an expected allocation or breakdown — for example, one expense line disaggregated across Program Services, Management, and Fundraising columns, or a list of component lines that sum to a total elsewhere on the statement. These represent expected differences, not genuine inconsistencies.
• IC group formation is more precise: each member of a group must have an individual reason to belong, reducing cases where unrelated occurrences were incorrectly grouped together.

Foot/Crossfoot validation accuracy

• The Foot/Crossfoot check now identifies more total rows across your statements and produces fewer incorrect corrections, including on statements where some formulas are genuinely wrong elsewhere on the same page.

Workpaper

• Uploads of large files are more resilient. When the initial processing request took too long to return a response, files could appear as failed even though processing was still running in the background. The upload now checks the actual processing status and updates automatically when it completes.

Task Overview

• The task overview page no longer fails to load when a category was deleted. If a task's assigned category had been removed from the engagement, opening that task would previously show an error. The page now opens normally and treats the task as uncategorized.

New Features

Client file downloads

Clients can now download all engagement files as a ZIP archive directly from the engagement page. Previously, clients could only access the Status Report — the files section was visible only to auditors. The new "Files → Download .zip" option gives clients access to uploaded evidence files and reference materials. Auditor-only documents are never included.

If any files cannot be packaged, the download will still complete and include a plain-text list of exactly what is missing, so clients always know what they received.

Improvements & Bug Fixes

Financial Statement Review

• Annotated PDF exports now open correctly in Adobe Acrobat Reader on Windows. Some exports were previously flagged as corrupt when opened in Acrobat (while opening fine in other PDF viewers). This is now resolved — exports work consistently across all major PDF viewers.

Stability

• Removing or replacing a datasource while AI is actively processing it no longer fails. Users previously encountered an error in this scenario that interrupted the operation; it now completes as expected.

Data Integrity

• Task title and description edits now always save. The first edit made to a task's title or description after opening the page was silently discarded — the interface showed success but the change was never stored. Subsequent edits on the same page load were fine. This is now fixed.
• Accounting policy changes persist immediately when navigating away. Adding, removing, or reordering accounting policies in the memo and then quickly leaving the page could silently drop those changes. Structural edits now save the moment you make them, so it is safe to navigate away or reload immediately after adding or deleting a policy.
• Accounting policy title and description edits save when you leave the field. Typing a policy title or description and then immediately switching pages could lose the edit. The field now saves as soon as focus leaves it, matching the behavior of all other memo fields.

Workpaper

• Large PDFs no longer show as failed while they are still processing. Files that go through a secondary OCR pass (used when initial text extraction quality is poor) can legitimately take longer to complete. Previously the page would declare these files failed too early, prompting unnecessary re-uploads. The check now correctly waits for the full processing window before reporting a failure.
• Files stuck in OCR processing surface as failed more reliably. When OCR was interrupted by a server issue, affected files could remain visually stuck indefinitely. They now correctly show as failed (red filename), and auditors can use Re-run OCR to reprocess them. Files that fail OCR three or more times in a row stay shown as failed until you manually trigger a re-run — automatic retries are suppressed to avoid a loop on files with consistently problematic content.
• Long AI prompt injected values are now readable. When a retrieval column prompt contained a long upstream value — for example, verbose findings from another column — the value was displayed as an oversized pill that was unreadable at paragraph length. Long injected values now render as a subtle inline text highlight, preserving readability at any length.
• Workpaper column group chevron now stays in sync. Collapsing or expanding a column group left the expand/collapse arrow pointing the wrong direction. The chevron now correctly reflects the current state.

Engagement Board

• Board no longer reloads unexpectedly. A spurious page reload could occur when certain browser extensions or session recording tools encountered a network error. The reload logic is now scoped to genuine app loading issues, so external network activity no longer triggers it.
• Select-all checkbox works consistently in grouped board view. When grouping the board by status, assignee, or another property, the column header's select-all checkbox was missing its expected behavior. Grouped and ungrouped views now behave the same way.

Memo

• Accounting policy reorder arrows are easier to click accurately. The move-up and move-down buttons in accounting policy rows were small and tightly packed next to the delete button, making them hard to target and occasionally triggering accidental deletes. The buttons now have larger hit areas with more spacing between them.
• The accounting policies toolbar + button now opens the type picker reliably. Clicking the + button in the accounting policies section of the memo was not opening the policy type picker in certain configurations. This is fixed.

Other

• Assigning a category to a task for the first time no longer causes a display error. Selecting a category on a task that had never had one assigned could temporarily break the category label on that task. This is now fixed.
• Empty task titles can be clicked to edit. A task title that had been cleared to empty had no visible click target, making it impossible to edit without workarounds. An empty title now shows a placeholder that keeps the field clickable.
• The status-rename announcement modal no longer blocks the rest of the page. While this modal was open, clicking anywhere outside it — including board columns and other page controls — was silently swallowed rather than dismissed or acted upon. Clicking outside the modal now closes it.
• CoAudit chat no longer shows duplicate controls in certain task chats. In some cases the chat panel could display two sets of send and attachment controls simultaneously. This is now fixed.

What's New

• Structured triage interview after automation runs: When an automation run finishes, CoAudit now opens an interactive interview instead of a block of text. The AI presents the top three to five findings as individual cards — each with concrete fix options and a "handle later" escape — and works through a root-cause sequence: are files in the right groups? are matches correct? are failures genuine or a workflow mismatch? Workpapers created from a task template get an additional note when repeated failures suggest the issue lives in the template rather than individual rows. The same interview runs when you ask CoAudit to "review my analysis," and it won't re-read data already visible in the run card above.

• Preview what the AI will write before you confirm: When CoAudit proposes writing a value to a protected workpaper cell, the approval card now shows the pending value — or Pass / Fail for a testing column — plus the AI's one-line reasoning, before you confirm. Batches show up to eight rows at a glance so you can review the recommendations rather than accepting them blind.

• Citations toggle now governs the AI's citation behavior in workbook responses: The Citations switch in the chat header controls how actively the AI cites workbook cell values. On — the AI treats citing every supported value as mandatory and finalizes citations before responding. Off — the AI cites workbook cells only when you explicitly ask. The ability to add citations is always available regardless of the setting.

Improvements

• Firm Library loads without freezing: Firms with very large libraries no longer experience the browser becoming unresponsive when opening the Library page. The list now renders only what is visible on screen, so the page is responsive immediately regardless of file count.

Bug Fixes

• Workpaper editor suggestion pickers no longer crash: The @ mention, / skill, and # assertion suggestion pickers no longer cause the workpaper editor to crash when document content is updated while a picker is open.

• Clear message when a document group name is already taken: Creating a document group with a name already in use on the workpaper now shows a specific "Name already in use" notice instead of a generic failure message.

• Sidebar updates immediately after joining or leaving an engagement: Adding or removing yourself from an engagement now refreshes the left sidebar list right away, without requiring a page reload.

• Workpaper cells save reliably when multiple changes land at the same time: Simultaneous edits to workpaper cells are now saved reliably — an edit no longer fails when another save is happening concurrently.

• Files with unusual characters process correctly: OCR output and AI responses that include uncommon text characters now save correctly instead of failing partway through processing.

• Client Organizations list is more resilient: The client list now handles brief connection interruptions gracefully, showing your clients instead of an error page.

• "Replace Rows" replaces "Replace Selection Data": The sampling split-button dropdown option has been renamed from "Replace Selection Data" to "Replace Rows" for clarity.

Bug Fixes

• CoAudit's activity feed now shows consistent, readable steps when helper agents are at work: CoAudit sometimes delegates parts of a task to specialized helper agents. Their activity now appears in the same clean format as the main agent's steps — showing actual file names, collapsing repeated actions into a single grouped count, and displaying a compact summary once the work is done. Tapping the summary expands a scrollable view of every step. Previously, this activity could show garbled internal codes instead of file names, repeat the same step dozens of times and flood the sidebar, or disappear from the chat entirely after a page reload.

• Cited workpaper cell references in Excel exports no longer fail on large workpapers: When CoAudit builds an Excel export that cites specific workpaper cells, it can now anchor those citations directly from the workpaper data it already scanned — without going back to re-read source documents a second time. This resolves a failure that previously blocked cited Excel exports when working across a large workpaper.

New in CoAudit

• Clickable citations in AI-produced Excel workbooks: When CoAudit builds an Excel workbook for you, individual cells can now include citations that link directly to the source document pages — the same click-to-open experience you get from a workpaper exported in Punchcard. Ask CoAudit to build you a workbook and it will wire up the citations automatically.
• CoAudit can now delegate complex tasks to worker agents: For requests that require reading many files and authoring detailed evidence across a large workpaper, CoAudit can now hand off that exploration to a focused worker — keeping your conversation responsive while the heavy lifting runs. This capability was previously available only to administrators; it is now part of every task-mode CoAudit chat.
• Filter workpaper rows by review state: CoAudit can now scope its reads and actions to only the rows you've flagged, validated, or haven't reviewed yet. Ask it to "look at only the flagged rows" or "re-run only unreviewed items" and it will target exactly that subset.
• Automation re-runs preserve your reviewed work: CoAudit now correctly understands that re-running a workpaper automation never overwrites cells you've already validated or flagged — so you can safely ask it to refresh the rest of the workpaper without worrying about losing your review work.
• Upgraded AI model: CoAudit chat now runs on an upgraded model with noticeably better accuracy on complex routing and multi-step tasks.

Chat & UI

• Drag folders and ZIP files into the chat: You can now drop a folder (or several folders at once), or a ZIP archive, directly into the CoAudit chat composer — all files inside are collected and uploaded as individual attachments. Previously only individual files could be dragged in.
• Scroll position stays put when the AI finishes responding: When the AI completes a reply, your reading position is no longer snapped to the bottom of the conversation. If you scrolled up to read an earlier message, you stay there. The scroll-to-bottom button becomes highlighted with a bouncing arrow to signal that a new reply is waiting — clicking it or scrolling down clears the indicator.
• Redesigned AI loading indicator: The indicator shown while the AI is generating a response is now a slim, solid bar pinned at the bottom of the conversation with an animated Punchcard mark and status text. It no longer lets transcript content bleed through behind it.

Requests

• Simplified request status labels: Client-facing request statuses are now Outstanding, Returned, Submitted, and Accepted. The button clients use to submit work is now labeled Submit Request; when a request is already submitted, it shows Return to Outstanding instead. Both auditors and clients will see a one-time announcement on the board summarizing the label changes side-by-side.

Bug Fixes

• Testing attribute name saves reliably on the memo page: Typing a name and then quickly reloading the page or closing the tab no longer silently discards the change — it now saves the moment you click away from the field.
• ZIP file previews work correctly in CoAudit chat: Files extracted from an uploaded ZIP now preview correctly in the document pane instead of showing a fetch error.
• PDF black-page recovery improved: Black or blank PDF pages are now detected and recovered from the moment a document opens, not only when navigating between pages.
• Collective workpaper columns compute correctly across multiple files: A column configured to produce one shared result across all files in a datasource group now generates a single synchronized value and applies it to every file, rather than running independently per file.
• Excel workbook delivery is more reliable: In rare cases, generating an Excel artifact could fail due to an internal step unrelated to the workbook's content. That step now fails gracefully so your workbook is still delivered.
• Task breadcrumb segments no longer collapse: Breadcrumb text in the task header now maintains a visible minimum width instead of occasionally collapsing to nothing.

Features

• Workpaper exports match your current sort order. Excel exports — Workpaper Only, Files + Workpaper, and Table Data Only — now output rows in the exact order displayed in the grid. If you sorted by balance largest-to-smallest before exporting, the downloaded file reflects that, with row numbers renumbered to match. Your sort is also saved automatically and restored when you navigate away and come back.

• Clickable source chips in the cell panel. When you open a cell's detail panel to view the prompt or formula CoAudit used, values that were drawn from other columns now appear as highlighted, clickable chips. Click a chip to jump directly to the source cell in the grid; a back arrow in the panel header lets you return to where you started. This makes it straightforward to trace exactly which upstream values shaped a given result.

• Skill mentions appear as pills in the main chat view. When you type an @[Skill name] mention in the standalone chat, it now renders as a blue pill — matching the visual treatment in the task chat sidebar — so it's immediately clear the skill was recognized and will be applied when you send the message.

Bug Fixes

• Engagement variable form stays open after a failed save. If you try to create a variable with a name that already exists, the form now stays open with your entries intact so you can correct and retry — instead of closing as though it had saved successfully.

• Bulk file uploads recover from temporary conflicts automatically. Uploading a large batch of files at once (common in payroll workflows) no longer surfaces intermittent "Upload Failed" errors when multiple files finish around the same time. Conflicts are resolved and retried automatically in the background.

• Large PDFs no longer trigger a false re-upload prompt. Files that take 10–15 minutes to extract — such as large payroll PDFs — sometimes showed a timeout warning and prompted a re-upload even though processing was still running and would have completed. That false alarm no longer occurs.

Features

• Import Deep Analysis questions into Skills: Each existing Deep Analysis question now has an "Import to Skill" button that opens the Skills editor pre-filled with the question's content and name. The option to add new Deep Analysis questions has been replaced with a link to the Skills library — existing questions continue to run, edit, and delete as before.

Reliability Fixes

• Files still processing after upload no longer appear as failed: When a large file takes longer than expected to finish text recognition, the workpaper now shows a "still processing" notice rather than marking it as failed. The status updates automatically once processing completes, so there is no need to re-upload.

• Failed saves now revert to the last confirmed value: Across status changes, memo fields, review states, task names, and descriptions, a save that doesn't go through now reliably restores the field to the last value the server confirmed, instead of leaving it stuck in the unsaved state.

• Switching tasks no longer mixes up edits between tasks: A save started on one task will no longer affect the task you navigate to afterward. Each save stays in the context of the task where it was made.

• Review status changes now update all related items immediately: Accepting or clearing a review state on one workpaper item in a linked group now immediately reflects on the other items in that group in the same view, without requiring a refresh.

• Narrative card review badges stay in sync: Review status badges on narrative cards now consistently reflect the latest confirmed state, even when a save and a concurrent change arrive at the same time.