What’s New in Parse: New Interface, Dynamic Object Detection Updates, and Workflow Improvements
This release is the most significant visual and structural update to Parse since launch. The interface has been rebuilt around how estimators actually move through a document, from initial upload to export. Navigation is faster, results are easier to act on, and the underlying analysis and extraction services have been restructured to support the new flows. Performance improvements across document processing are also included in this release.
Rebuilt Interface: A New Way to Move Through Construction Document Analysis
What changed:
The analysis view now uses a tabbed layout that separates dynamic object results, page-level navigation, risk and scope views, and AI chat into distinct, purpose-built panels. Identified Items at the top of the view shows the total count of everything extracted from the document at a glance.
Why it matters:
Pre-construction document analysis covers a lot of ground. The previous interface asked estimators to hold too much in their heads at once. The tabbed structure means you are always looking at one view, one job. Move between tabs as the workflow demands without losing your place.
Dynamic Object Detection: Configuration, Quantities, and Page-Level Review
Dynamic object detection has received meaningful adjustments in both how objects are configured and how results are presented.
What changed:
The Configure Analysis panel now shows the full list of detected dynamic objects with individual toggles. You can turn specific object types on or off before running or re-running analysis. From the Dynamic Objects tab, each detected object can be expanded to show additional detail and its highlighted location directly on the document. The new Page tab lets you navigate all dynamic objects by the page they appear on.
Why it matters:
On a complex set of drawings, dynamic object detection can surface dozens of object types across hundreds of pages. The ability to toggle specific objects in configuration, expand individual results, and navigate by page means estimators work through results in a structured way rather than scrolling through a flat list. For multi-trade bids, filtering down to the object types that matter for the scope in front of you changes how fast you can move.
Text Documents: CSI Division Default View and Line Item Navigation
The analysis view for text-based documents, including specs, RFPs, and scope documents, now defaults to a CSI division-organized layout.
What changed:
Results for spec books and RFP documents now display organized by CSI division by default. Each division can be expanded to reveal individual line items. Clicking a line item navigates directly to the source page in the document. The Page tab shows all results organized by page number, and the Risks tab organizes results by scoping category for a different angle on the same data.
Why it matters:
CSI division structure is how estimators think about scope. Defaulting to that view in construction document analysis means less sorting and reorganizing after extraction. The source-link navigation, clicking a line item and landing on the exact page it came from, keeps the verification step fast and auditable without context switching.
Export: Document-Level and Page-Level Options with Format Selection
What changed:
Export is now split into two distinct options: Export Document and Export Page. From the Export Page menu, you can select your preferred format, including Page Excel Report. Items are selected for export using checkboxes at the category or line item level, so you control exactly what goes out.
Why it matters:
Not every export needs the full document. For estimators pulling specific trade quantities or materials specs into a bid package, page-level export in Excel format means Parse outputs land directly where they need to go without cleanup.
Document Versions and Comparison View
What changed:
Documents now include a Versions tab where additional versions can be uploaded and named. Each new version processes automatically using the same configuration as the original. The Comparison View shows differences between versions in a side-by-side list, with each result linking to the highlighted change in the document. Page navigation within the comparison view lets you move through the full set of changes systematically.
Why it matters:
Addenda and revised drawings are a constant in commercial pre-con. The version and comparison workflow means your team is never manually hunting for what changed between a drawing set and its revision. Differences surface in a list, each one linked to its source location in both documents.
AI Chat: Page Configuration and Hyperlinked Results
What changed:
AI Chat is now available as a dedicated tab within every document analysis view, in addition to the project-level chat. Within document-level chat, you can configure the number of pages the AI has access to for more focused results on large documents. All AI chat results are hyperlinks that navigate directly to the referenced page in the document.
Why it matters:
Limiting AI chat to a relevant page range significantly improves answer quality on large plan sets where full-document context can dilute precision. Hyperlinked results close the loop between the AI answer and the source, which is the auditability requirement that makes construction document analysis trustworthy in a bid context.
Data Structure and Performance Updates
What changed:
Underlying data structures in the Analysis and Extraction Services have been updated to support the new interface and navigation model. Performance improvements reduce processing time across document types.nt.
Why it matters:
These changes are largely invisible but materially improve the reliability and speed of everything built on top of them. Faster processing, more consistent extraction, and a data model that supports the new UI without workarounds.
These updates are live now!
Upload a document, move through the new tabbed analysis view, configure your dynamic object detection, and run a comparison between two versions of a drawing set. Tell us what the new interface gets right and where it needs work.
Send us your worst document
