Extracting data from equity incentive plans
Companies grant stock awards under a plan the board adopts, each drafted in outside counsel's own structure. Proxy advisers score every plan put to a stockholder vote, reading it from SEC exhibits and proxy statements. Acquirers review the target's plans in diligence, as PDFs in a data room.
Analysts need the share reserve, the counting ratio for full-value awards, the director award cap and the change-in-control trigger. A dilution model combines those with outstanding awards to measure overhang, and a deal model costs the awards that accelerate on a sale.
What makes equity incentive plans hard
Figures buried in legal prose
The share-counting rule can state three ratios in one sentence, each tied to the dates it applies between, inside a paragraph of drafting.
Each restatement moves the numbers
Companies restate a plan whenever they add shares, and each version can change the counting ratio and the award caps.
Quantities written as words
Vesting periods and option terms are often written out, as in "one year" and "ten years", and a dilution model needs them as numbers.
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 equity incentive plan, end to end
About this document: The Stock and Incentive Compensation Plan of Alcoa Corporation, as amended and restated, published on the company's equity plan page. Eighteen portrait pages of single-column legal text, from definitions to the plan term. Page 5 carries the director award cap and Section 4(a), the share reserve. The file is published as downloaded.
Processing objective: The objective is to pull the size of the share pool, and how fast awards draw it down, out of running legal prose. Each figure sits inside a sentence that carries several others, so its box has to land on the one line that prints it.
Demonstration fields: Shares reserved, Incentive stock option limit, Full-value award counting ratio, Ratio effective from, Director annual award cap
The document
4 of 18 pages
Page 1
Page 5shown below
Page 6
Page 12
Where each value came from
share_reserve.total_shares_reserved
{
"value": 38000000,
"page": 5,
"range": {
"end": 19918,
"start": 19823
},
"box": {
"xmin": 0.17258,
"ymin": 0.77692,
"xmax": 0.88298,
"ymax": 0.79138
}
}share_reserve.incentive_stock_option_limit
{
"value": 38000000,
"page": 5,
"range": {
"end": 20291,
"start": 20202
},
"box": {
"xmin": 0.17205,
"ymin": 0.83507,
"xmax": 0.85753,
"ymax": 0.84967
}
}share_reserve.full_value_award_ratios[2].shares_counted
{
"value": 1.69,
"page": 5,
"range": {
"end": 20482,
"start": 20388
},
"box": {
"xmin": 0.17188,
"ymin": 0.86415,
"xmax": 0.86311,
"ymax": 0.87897
}
}share_reserve.full_value_award_ratios[2].period_start
{
"value": "May 6, 2026",
"page": 5,
"range": {
"end": 20573,
"start": 20483
},
"box": {
"xmin": 0.17162,
"ymin": 0.8787,
"xmax": 0.8403,
"ymax": 0.89343
}
}director_annual_award_limit
{
"value": 750000,
"page": 5,
"range": {
"end": 18461,
"start": 18366
},
"box": {
"xmin": 0.11479,
"ymin": 0.49934,
"xmax": 0.82456,
"ymax": 0.51413
}
}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.
