Extracting data from vehicle schedules
Commercial fleet operators and their brokers list every insured unit in a vehicle schedule, usually a spreadsheet in the insured's own columns. Carrier underwriters receive one with each new submission and renewal, a row for every unit. It arrives as an Excel file or PDF export, attached to an email or a public-entity insurance RFP.
Underwriting needs each unit's year, make, model, VIN, body type, stated value and garaging location. A rating system classifies every unit by type and territory and prices physical damage on the stated value. The unit list then carries onto the policy's auto schedule.
What makes vehicle schedules hard
Every insured builds its own columns
The value column is headed Agreed Value on one fleet's spreadsheet and Stated Amount or Cost New on another, in whatever position the insured chose.
Units grouped under header rows
A fleet list often groups its units under full-width header rows for each location or department, and a group can continue onto the next page.
Equipment listed among the vehicles
A schedule lists trailers, watercraft and portable equipment alongside road vehicles, so a value column can hold a coverage note where an amount would print.
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 vehicle schedule, end to end
About this document: Addendum 1 to a property and casualty insurance RFP from the County of Louisa, Virginia. The county published it so bidding insurers could price its fire and EMS program. Its 4 pages open with questions and answers. Pages 3 and 4 carry the vehicle schedule: 86 units in a nine-column table under twelve section headers. Page 3 is featured.
Processing objective: The objective is to capture one fleet line the way an underwriter rates it, from the department header down to the unit's insured value and its garaging station. The department name prints once, in the header row above, and covers every unit beneath it.
Demonstration fields: Department section, Body type, VIN, Agreed value, Station ID
The document
4 of 4 pages
Page 1
Page 2
Page 3shown below
Page 4
Where each value came from
locations[0].section_name
{
"value": "Louisa Vol. Fire Dept.",
"page": 3,
"range": {
"end": 2103,
"start": 2081
},
"box": {
"xmin": 0.08308,
"ymin": 0.09506,
"xmax": 0.83516,
"ymax": 0.10679
}
}locations[0].vehicles[0].body_type
{
"value": "Pumper",
"page": 3,
"range": {
"end": 2175,
"start": 2169
},
"box": {
"xmin": 0.34147,
"ymin": 0.10678,
"xmax": 0.48551,
"ymax": 0.11787
}
}locations[0].vehicles[0].vin
{
"value": "4PICV01E5BA011783",
"page": 3,
"range": {
"end": 2201,
"start": 2184
},
"box": {
"xmin": 0.48551,
"ymin": 0.10677,
"xmax": 0.63411,
"ymax": 0.11783
}
}locations[0].vehicles[0].agreed_value
{
"value": 580000,
"page": 3,
"range": {
"end": 2217,
"start": 2210
},
"box": {
"xmin": 0.63411,
"ymin": 0.10677,
"xmax": 0.73475,
"ymax": 0.11779
}
}locations[0].vehicles[0].unique_id
{
"value": "FS1",
"page": 3,
"range": {
"end": 2229,
"start": 2226
},
"box": {
"xmin": 0.73474,
"ymin": 0.10676,
"xmax": 0.80443,
"ymax": 0.11776
}
}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.
