Document Recognition

Document Intelligence

Turning the paperwork every construction project generates — contracts, job forms, change orders, pay applications — into structured, searchable data automatically, instead of the manual review it takes today. Coming Soon — nothing on this page is live yet.

How It Will Work

Document layout and field extraction (OCR plus a layout-aware model, since construction paperwork is rarely plain text — tables, continuation sheets, and stamped signatures all carry meaning) turns a scanned or exported PDF into structured fields: parties, dates, dollar amounts, line items, and references to other documents (a change order citing its original contract, a pay application citing a schedule of values). Classification comes first — knowing a document is a G702 pay application versus a daily field report determines which fields even apply — before any extraction happens.

File Formats

Built to Take Whatever a Jobsite Actually Produces

Construction paperwork doesn't arrive as clean, uniform text — it's scanned, photographed, and exported from half a dozen different systems. The plan is to take multiple input formats natively rather than forcing everything through a single upload path first.

PDF
PDF (digital or scanned)
The primary format construction paperwork actually arrives in — including scanned, image-only PDFs with no embedded text layer.
JPG / PNG
Photographed pages
A daily field report or a stamped drawing shot on a phone at the jobsite, not just an office-scanned document.
TIFF
Multi-page TIFF
The format most large-format scanners and older document-management systems still export by default.
DOCX / XLSX
Native Word & Excel
Change orders and pay application continuation sheets that were never printed at all — read directly, not via a screenshot.
CSV
Structured exports
A schedule of values or a log already exported from another system — ingested directly instead of re-typed.

Document Types

What It Will Read

Owner-Contractor Agreements: Standard-form construction contracts (e.g. AIA A101/A102-style lump-sum or cost-plus agreements) — scope, price, schedule, and retainage terms read directly from the executed document.
DOB Job Forms: PW1 (Plan/Work Application), TR1 (Statement of Special Inspections), and the rest of the DOB filing paperwork — see the Permit Guide for what each form covers.
Change Orders: Standard-form change orders (e.g. AIA G701-style) — scope change, cost impact, and schedule impact extracted and tied back to the original contract line item.
Pay Applications: Application and Certificate for Payment packages (e.g. AIA G702/G703-style continuation sheets) — line-item percent complete, retainage held, and amount due read automatically instead of re-keyed.
RFIs & Submittals: Requests for Information and submittal logs — tracked, dated, and tied to the spec section or drawing they reference.
Daily Field Reports: Superintendent daily logs — weather, crew counts, equipment on site, and work performed — structured into a queryable timeline instead of a folder of PDFs.

Integrations

Read Where Documents Already Live

The plan is to connect to a project's existing document storage, not require a new upload workflow — documents get read in place.

Dropbox
Point Document Recognition at a project's existing Dropbox folder structure — no separate upload workflow, no re-organizing files to fit a new tool.
Coming Soon
Microsoft SharePoint
Read directly from the document libraries most GCs and owners already use for project records, permitting, and closeout documentation.
Coming Soon
Google Drive
For teams standardized on Google Workspace — same read-in-place approach as Dropbox and SharePoint, no migration required.
Coming Soon
Box
Enterprise-grade document storage, common on larger GC and owner-side teams with stricter access-control requirements.
Coming Soon

See the full product roadmap on Our Technology, or the DOB forms this will start with on the Permit Guide.