Making Board Disclosures AI Legible

Making Board Disclosures AI Legible
Download PDF

Introduction

In 2010, we launched Labrador in the United States with a simple idea: corporate disclosure should be easy to read and understand for investors and other stakeholders. Back then, that idea was rare enough that we were the only firm built around it. Sixteen years later, thanks to engaged reporters like you, annual meeting and sustainability disclosures, and other investor and stakeholder communications have improved enormously. Clear, transparent disclosure has become the norm, and at Labrador we are proud to have played a part in that evolution.

In 2026, our mission to make your disclosures easy to read and understand hasn’t changed… but your readers have.

Your investors are still reading your disclosures. But increasingly, so is an AI model, deployed by an institutional investor, a proxy advisor, or an analyst, to parse your governance disclosures to find the same answers sought by the human readers.

This rapid evolution in your readers led us to ask, “if a human can read a page and understand it in seconds, can a machine do the same”? We started to find answers in 2022, when we first piloted AI to accelerate our analysis of disclosures for the Transparency Awards. That work taught us a lot, not just about what AI is remarkably good at extracting from a well-built document, but also about what it misses and why. Icons with no accompanying text. Charts with data trapped inside an image. Skill matrices that a human eye reads instantly but a machine can’t parse.

We now have incorporated what we learned into our advisory practice and the design system we use on every engagement. Today, when we help a client prepare disclosure, we’re thinking about the human reader and the algorithm running in the background. The goal is the same content and design telling the same story clearly to both.

That’s what this Thought Piece is about. In the following examples, we are sharing some of the practical, specific things we’ve learned, the kind of tips you can easily apply to board and governance disclosures in your next proxy statement to make them visually compelling and machine-readable. Every idea proposed here is something we build into our own project processes and is EDGAR compatible.

If you’d like a second opinion on where your own disclosures stand beyond these governance examples, we’d be glad to help. We’ll run a no-fee AI-readiness audit on your materials and hand you clear, actionable ideas with no strings attached. Just get in touch!

It’s been a professional thrill to accompany corporations toward better disclosure and stronger outcomes over these sixteen years, and we couldn’t be more excited about what comes next.

Making Proxy Disclosures Readable by AI 1

A complimentary review of
the AI Legibility of your disclosures

In January 2026, J.P. Morgan Asset Management replaced ISS and Glass Lewis with an in-house AI tool for its US proxy voting. In the Center for Audit Quality’s July 2026 survey, only 2% of institutional investors said they do not use AI when reviewing company filings. Those tools read a document’s text and data layer. They do not read pictures.

What we do. We review your proxy statement, annual report, Form 10-K or sustainability report in the four ways an automated system does and compare the results. Where the methods disagree with one another, that disagreement is the finding. We work only from what is already public: no data, no system access, no time from your team.

What you receive. A written report: what is already clear to AI, what is not, where the document could lead an automated reader to the wrong answer, and a prioritized table of fixes, with precise instructions on how to implement.

The review is complimentary and carries no obligation. Reach out to Iain at poole.i@labrador-company.com and we will do the rest.

Making Proxy Disclosures Readable by AI 2

Director Nominees Summary Table

Many proxy statements feature an at-a-glance table listing each nominee’s biographical information, independence, and committee memberships.

Making Proxy Disclosures Readable by AI 3
Making Proxy Disclosures Readable by AI 4
Making Proxy Disclosures Readable by AI 5

Board Composition, Tenure, and Diversity

A chart appearing just after the nominees table, intended to summarize the tenure distribution and diversity composition of the Board. 

Making Proxy Disclosures Readable by AI 6
Making Proxy Disclosures Readable by AI 7
The chart and image presentations are still a .jpg files because EDGAR leaves no alternative. Every value it displays also sits beside it as fully-searchable live text. Nothing is sacrificed visually, the graphic is an effective anchor for the reader skimming the page, and both human and machine readers can grasp the numbers.
Making Proxy Disclosures Readable by AI 8
Because the legend is real HTML text rather than pixels, EDGAR’s full-text search indexes it, a screen reader reads it aloud, an analyst can copy it straight into a spreadsheet, and an AI summarizing the filing quotes the actual figure instead of guessing at an image of a slice or block.

Board Skills Matrix ​

Many proxy statements feature an at-a-glance table listing each nominee’s biographical information, independence, and committee memberships.

Making Proxy Disclosures Readable by AI 9

Without a clear understanding of either names or datapoints, the matrix isn’t useful. Although in some cases director biographies list skills individually, mitigating this issue, a machine-readable matrix conveys its meaning through row labels, the column headers, and marks, and if one (or more) are unreadable, the remaining one or two become worthless. 

Making Proxy Disclosures Readable by AI 10
  • The “After” matrix is set as a true HTML <table>, so intersections survive. Row label, column header and cell stay bound together, which makes the relationship between the data clear, preserving the communication objective of the matrix.
  • Icons, although unreadable to AI, earn their place because they make a page easier and more pleasant for the human reader.

Director Biographies

The individual biography pages for each nominee, combining a rationale for nomination, career history, and a “key skills and expertise” narrative.

Making Proxy Disclosures Readable by AI 11
Making Proxy Disclosures Readable by AI 12

The result

  • Every fact resolves into a structured record — and, unlike prose, the same record can be built for all eleven directors and compared across companies.
  • A machine can now compute tenure, career length and how recent the operating experience is. None of that was possible before.
Download PDF
PARTAGER :
logo r1
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.