Extract data from proxy statements

A proxy statement is the document a public company sends its shareholders so they can vote without attending the annual meeting in person. Filed with the SEC as a DEF 14A, it is the governance counterpart to the financial filings: where a 10-K reports the numbers, the proxy asks the owners to elect the board, approve executive pay, and ratify the auditor. Governance analysts, proxy advisors, and institutional stewardship teams all read the same document for the same facts: when the meeting is held, the record date that fixes who may vote, the slate of director nominees and whether each is independent, the executive compensation the board is asking shareholders to endorse, and the specific proposals on the ballot with the board's recommendation on each. Because the vote is time-bound, the meeting date and the record date drive the whole workflow. Substance sits in a few dense tables, not in the narrative. The Summary Compensation Table breaks each named executive officer's pay into base salary, bonus, stock awards, non-equity incentive, and a total that has to foot. Director nominees are listed with a position, committee memberships such as Audit or Compensation, and an independence classification under the exchange listing rules. Each proposal carries a number, a type such as Director Election or Say-on-Pay, a plain-text description, and a board recommendation of For or Against. After the meeting, the 8-K vote results report votes for, against, and abstained, plus broker non-votes, and whether the item passed. Stewardship teams reconcile the recommendation against the outcome across hundreds of filings in a proxy season. Talonic reads the proxy statement into a structured record that keeps each table intact. A DEF 14A from Cascade Semiconductor Corporation (CIK 0001472051, ticker CSMC), for an annual meeting on 2026-05-14 with a record date of 2026-03-18, returns its three director nominees, a Summary Compensation Table where chief executive Raymond Beck shows a base salary of $950,000, stock awards of $4,200,000, and a total of $7,000,000 that foots, and three proposals including the ratification of Ernst & Young LLP, so a governance data set is populated from the filing rather than keyed by hand.

What gets extracted from proxy statements

How extraction works for proxy statements

Proxy statements are filed as DEF 14A documents with the SEC and reach an analyst as EDGAR PDFs, print renders, or vendor feeds, and the governance data lives in tables laid out differently by every issuer. Talonic classifies the filing and maps each value to the proxy field set held in the Field Registry, which separates the company and meeting header, the board nominee list, the Summary Compensation Table, the proposals, and the voting results into distinct blocks rather than one flat page. Meeting, record, and distribution dates are parsed to ISO 8601, each nominee keeps its position, committee memberships, and independence status, and every compensation row keeps its salary, bonus, stock awards, non-equity incentive, and total, which is checked to foot. Each proposal keeps its number, type, description, and board recommendation, and where vote results are present the votes for, against, abstained, and broker non-votes are typed as numbers. Per-value confidence and a pixel-region pointer under DIN SPEC 91491 let a stewardship team verify a compensation figure or a recommendation against the source. The extraction structures what the proxy states and offers no voting or investment advice.

Sample extraction

A DEF 14A proxy statement for an annual shareholder meeting

{
  "document_number": "DEF 14A",
  "document_date": "2026-04-02",
  "company.name": "Cascade Semiconductor Corporation",
  "company.cik": "0001472051",
  "company.ticker": "CSMC",
  "company.state_incorporation": "Delaware",
  "meeting.date": "2026-05-14",
  "meeting.location": "Virtual meeting via live webcast",
  "meeting.time": "9:00 a.m. Pacific Time",
  "record_date": "2026-03-18",
  "distribution_date": "2026-04-02",
  "board_members": [
    {
      "board_member.name": "Diane Okafor",
      "board_member.position": "Chair, Independent Director",
      "board_member.committee_membership": [
        "Nominating"
      ],
      "board_member.independence_status": "independent"
    },
    {
      "board_member.name": "Raymond Beck",
      "board_member.position": "Chief Executive Officer, Director",
      "board_member.committee_membership": [],
      "board_member.independence_status": "non-independent"
    },
    {
      "board_member.name": "Priya Nandakumar",
      "board_member.position": "Independent Director",
      "board_member.committee_membership": [
        "Audit",
        "Compensation"
      ],
      "board_member.independence_status": "independent"
    }
  ],
  "compensation_summary": [
    {
      "executive_officer.name": "Raymond Beck",
      "compensation.annual_salary": 950000,
      "compensation.bonus": 0,
      "compensation.stock_awards": 4200000,
      "compensation.non_equity_incentive": 1850000,
      "compensation.total_compensation": 7000000
    }
  ],
  "proposals": [
    {
      "proposal.number": "1",
      "proposal.type": "Director Election",
      "proposal.title": "Election of three directors",
      "proposal.recommendation": "For"
    },
    {
      "proposal.number": "2",
      "proposal.type": "Say-on-Pay",
      "proposal.title": "Advisory vote on executive compensation",
      "proposal.recommendation": "For"
    },
    {
      "proposal.number": "3",
      "proposal.type": "Auditor Ratification",
      "proposal.title": "Ratification of Ernst & Young LLP as independent auditor",
      "proposal.recommendation": "For"
    }
  ],
  "voting_results": [
    {
      "voting_results.proposal_number": "2",
      "voting_results.votes_for": 42100000,
      "voting_results.votes_against": 3200000,
      "voting_results.votes_abstain": 400000,
      "voting_results.broker_non_votes": 1800000,
      "voting_results.result": "Passed"
    }
  ]
}

Frequently asked