Extracting data from G703 continuation sheets
General contractors attach a G703 continuation sheet to every monthly pay application, and many build their own spreadsheet edition with relabeled columns. Owners and construction lenders review one for each active project, every billing period. The sheets arrive as emailed PDFs, Excel exports or downloads from project management portals.
Draw review needs each line's scheduled value, change orders, work completed to date and retainage. A construction accounting or draw management system checks that the lines sum to the G702 summary, then releases payment less the retained amount.
What makes G703 continuation sheets hard
Every column depends on another
Each row chains its figures: previous work, this period and stored materials sum to total completed, and the balance to finish follows.
Contractors print their own editions
Many contractors rebuild the form in a spreadsheet, adding change order columns and relettering the rest, so column positions differ by contractor.
Schedules of values run long
A large job's schedule of values lists every trade by CSI division, plus allowance and contingency lines, across several continuation pages.
Parse, extract, verify
Parse
Parse turns the document into reading-order Markdown and a hierarchy of blocks (text, tables, figures), and grounds each block to its exact location on the page.
Extract
You define the fields in a JSON schema, and Extract returns their values in a consistent, structured format, each linked back to where it appears in the Markdown.
Verify
Each value links back to where it appears, down to the character range in the Markdown, and the page and bounding box in the source document, so every result can be verified.
One AIA G703 continuation sheet, end to end
About this document: A pay application for the remodel of city center buildings in Lake Elmo, Minnesota, posted in a city council packet. Two landscape pages: a G702-style application summary, then a continuation sheet the construction manager built as its own spreadsheet edition of the G703 layout. Page 2 is featured. Signatures on page 1 were replaced with stand-ins and a notary stamp was removed.
Processing objective: The objective is to follow one schedule of values line through a change order and read the job's total completed to date. The HVAC line reconciles: a scheduled value of 962,000.00 plus 154,703.00 in change orders gives the revised 1,116,703.00.
Demonstration fields: Period ending, HVAC and Controls scheduled value, HVAC and Controls change orders, HVAC and Controls revised value, Total completed and stored to date
The document
2 of 2 pages
Page 1
Page 2shown below
Where each value came from
application.period_to
{
"value": "10/31/2022",
"page": 2,
"range": {
"end": 3683,
"start": 3447
},
"box": {
"xmin": 0.63273,
"ymin": 0.06164,
"xmax": 0.77492,
"ymax": 0.09133
}
}schedule_of_values[26].scheduled_value
{
"value": 962000,
"page": 2,
"range": {
"end": 9535,
"start": 9525
},
"box": {
"xmin": 0.2386,
"ymin": 0.66539,
"xmax": 0.30338,
"ymax": 0.68094
}
}schedule_of_values[26].change_orders
{
"value": 154703,
"page": 2,
"range": {
"end": 9554,
"start": 9544
},
"box": {
"xmin": 0.30338,
"ymin": 0.66538,
"xmax": 0.3647,
"ymax": 0.68095
}
}schedule_of_values[26].revised_scheduled_value
{
"value": 1116703,
"page": 2,
"range": {
"end": 9575,
"start": 9563
},
"box": {
"xmin": 0.3647,
"ymin": 0.66537,
"xmax": 0.43086,
"ymax": 0.68095
}
}totals.total_completed_and_stored_to_date
{
"value": 6849725.77,
"page": 2,
"range": {
"end": 10829,
"start": 10817
},
"box": {
"xmin": 0.55336,
"ymin": 0.75813,
"xmax": 0.61299,
"ymax": 0.77307
}
}How to try it yourself
Everything on this page is reproducible. The sample, its schema, and the Parse and Extract API outputs are published. You can repeat the same two calls from whichever surface you already build on.
- Claude Skilldescribe the job; the agent writes it
- CLIno code at all
- MCPexpose ADE docs as an agent tool
- REST APIany language
- Pythonlandingai-ade
- TypeScriptlandingai-ade
Documents into data
Accurate
The best accuracy on the difficult documents other tools get wrong, such as messy scans, complicated tables, checkboxes and handwriting.
Auditable
Every answer cites its sources. Values come back with an audit trail (page, bounding box, character range) so a human can review them.
Agentic
Applies multi-step reasoning, and plans before it acts. Handles documents it has never seen before and adapts to variations.
API-first
Composable APIs that drop into any tech stack and can be called from anywhere. Easy to build on and embed in your pipelines and products.
Affordable
Cost-competitive rates with complete price transparency. Spend controls before you run, and every job reports its exact cost.
Enterprise-grade
Built for production and scale from day one. Fully observable and monitored, running securely in our cloud or yours.
Built for high-volume, high-throughput production
- Zero data retention
- Priority rate limits
- Enhanced support
- VPC and on-prem deployment
Available on Team and Enterprise plans. VPC and on-prem deployment require an Enterprise plan.
