September 29
This release builds upon fileAI’s enterprise-ready authentication, new ways to manage Excel files, stronger audit controls with Citations and Validations, and deeper integration with Cloud storage systems.
New features
- Microsoft Single Sign-On (SSO) – Full support across web and mobile, allowing organizations to manage secure logins and onboarding through their existing Microsoft identity systems.
- Export to OneDrive & SharePoint – First release of direct export options, enabling customers to send processed documents directly into OneDrive or SharePoint without extra manual steps.
- Excel Splitting for Multi-Tab Files – Multi-tab Excel spreadsheets can now be split into unique files, making it easier to organize, process, and share large workbooks.
- Citations & Validations (Human-in-the-Loop) – Users can now view source citations alongside extracted values and validate or correct them directly in the UI. This builds trust in AI results, improves auditability, and gives teams more control over data quality.
Improvements
- Expanded Citations & Validations – Audit and review features have been extended to include more file types, such as tables in PDFs. Workspace admins can now toggle this feature at the project level for greater flexibility.
- Original File Handling – Downloaded files now retain their original file names, ensuring better traceability between source documents and processed outputs.
Fixes
- Corrected Safari compatibility for report generation.
- Restored missing invoice buttons on billing for self-serve users.
September 15
This release brings more transparency in extracted data, stronger workspace controls, and expanded file handling options to make document management more flexible and reliable.
New features
- Citation & validation transparency – Extracted data in JSON now includes source citations, with the option for manual override. Contact your account manager to enable this feature.
- Self-serve plan auto-renewal – Subscription plans now renew automatically, preventing service disruption.
- Download original files as PDF – From the File Details page, you can now download the original document as a PDF for easier archiving and compliance.
Improvements
- UI upgrades – Smarter table handling for large datasets and smoother list view scrolling.
- Folder handling – Better alignment between manual, automatic, and import folder creation for consistent organization.
- Microsoft Entra login – Expanded login support for broader authentication options.
- Excel support – Improved compatibility and performance when processing large or complex spreadsheets.
Fixes
- Fixed display and routing glitches across Browse and All Files tabs.
- Corrected issue where clicking a file still in “created” status showed a “document not found” screen.
- Resolved issue where prompting in AI Schema triggered unending reprocessing.
- Improved List View in All Files/Folders to prevent constant jumping of data.
- Fixed errors when changing operator from AI to fuzzy.
September 1
This release introduces powerful new AI-driven capabilities, smarter document organization and summaries, and deeper integration with Microsoft Entra and Xero. Alongside these, we’ve delivered important performance improvements and fixes to make document processing faster, more reliable, and easier to manage at scale.
New features
- Summary Pages for Uploads – Each upload now automatically generates a summary page per project space across web and mobile, giving users quick insights without digging into raw files.
- Microsoft Entra Integration – Improved sign-in and workspace management with Entra, including streamlined consent flows and error handling.
- AI Reconcile with NLP chat functionality – Launched Natural Language Processing (NLP) chat agent on web and backend, enabling users to ask questions in plain English and get instant data insights.
- Schema Locking – Added a schema locking mechanism to prevent accidental changes and improve data consistency.
- Xero Auto-Export Rule Creation – Enabled auto-export and auto-rule creation for Xero, simplifying finance workflows.
Improvements
- Processing Flow – Refined backend logic for the “Processing > Validating” flow, ensuring smoother handling of document states.
- Subfolder Performance – Reduced delays in trigger events for subfolder creation and Match & Compare.
- AI Reconcile Enhancements – Improved handling of table views and column embeddings for more accurate query results.
- Debugging & Logging – Refactored log formats and added tracing for clearer debugging across systems.
- NLP Exports – Added a download button for NLP result screens to simplify sharing and reporting.
Fixes
- Fixed issue where adding a project worked initially but failed to save.
- Corrected sync issues in the document directory.
- Fixed upload progress percentages not updating for master users.
- Restored missing fields in Excel exports for certain schemas.
- Corrected display errors when more than two files were present.
- Fixed file deletion errors showing “No file” when files existed.
- Corrected supplier order item misalignment in subfolders.
- Stopped deleted files from appearing in search results.
Infrastructure
- Upgraded orchestration tracing between services for better reliability.
- Improved logging for AI pipelines.
August 18
New features
- Manual file upload – You can now manually add files from Drive into subfolders.
- Export option – Added the ability to include the original document as an attachment when exporting to Xero.
- Translation support – Preserved translations when downloading documents.
- Auditability – Citations and validations are now displayed for extracted values.
- Fuzzy match thresholds – Introduced thresholds for fuzzy matching, giving users more control and accuracy when working with similar values.
Improvements
- Match & Compare – Improved backward compatibility for older versions.
- Schema editing – Added more flexibility in key identifier operators.
- AI model – Exposed a new AI model for document processing, Beethoven_ENG_O5.8.
Fixes
- Fixed issues where translated documents were not persisting across views and downloads.
- Corrected errors with routing rules and schema changes.
- Fixed duplicate documents and duplicate accounts appearing in some views.
- Resolved display glitches showing “No rows to show.”
Infrastructure
- Migrated databases to AWS Aurora for improved reliability.
- Backend refactoring for faster, more stable Match & Compare operations.