01What happens.
A board change is announced once and the previous board is recorded thousands of times: in annual reports, presentations, profiles, interviews, news coverage and third-party directories. A machine weighing the record finds the old names in the majority of it, and finds nothing in most of those documents to say the people have gone.
02Why it matters.
If you are the company secretary or the head of investor relations, the leadership of the company is the first fact anyone establishes before they act, and it is now established by asking AI. A wrong answer is a small embarrassment when a letter goes to the wrong person and a larger one when a proxy adviser, an activist or a journalist builds an argument on who is accountable for what 'do the board have any control of the basics here...'.
Outdated facts persist in AI descriptions of companies after the facts have changed, and leadership is the fact that changes most visibly and is corrected least thoroughly across the record.
03The case.
This illustrative scenario is written from the mechanism; no company is identified.
A listed company's chair retires at the annual general meeting after 9 years. A successor is appointed and announced. Four mnths. later the chief executive steps down and a new chief executive is named. Both changes are announced through the regulatory news service, the investor site's board page is updated the same day, and the new leadership is presented at the next results.
The record of the previous leadership stays where it was. There are 9 annual reports with the former chair's statement at the front, hundreds of pages of presentations and transcripts, interviews and profiles across the financial press, the chair's biography on two industry bodies' sites, and third-party company directories that refresh on their own schedules. The former chief executive's name sits on 6 years of the same material, and on the deck used at the last capital markets day.
An investor asks AI who chairs the company and who runs it. The answer names the former chair and the former chief executive, drawn from the weight of the record.
The board page, which is correct, is one page against 9 years of documents, and none of those documents says the people have left. The record is not wrong; it is old, and the machine cannot tell the difference. A shareholder letter arrives addressed to the former chair.
A proxy adviser's draft report names the former chief executive as accountable for the year's performance. A journalist's piece describes a leadership that no longer exists. Each is corrected by the company one at a time, and each correction is met with the same question from the person corrected: why does AI say otherwise?
The company had announced the changes properly and updated the page that mattered. What it could not see was that the page that mattered was outnumbered by the facts! and that nothing in the record carried the marking a machine reads as superseded.
04Exposure.
The board answers for how the company is represented, and the identity of the board is the first element of that representation.
A governance picture that is wrong at the first question undermines the account of accountability the company offers. The correction, issued repeatedly, reads as a company that cannot keep its own leadership straight, at the moment it is asking investors to trust a new one.
The cost is the correction and the correspondence. The cost that lasts is the credibility of the company's governance disclosures with the investors, proxy advisers and journalists who form a view of the board from what AI tells them, before they read anything the company published.
05What AI Readiness changes.
What the company cannot see from the inside is how far the old names outweigh the new ones across the record. AAAnow reads the company's digital landscape as AI reads it and evidences which documents AI is drawing on for the leadership question, on the company's own properties and beyond.
With that evidence the company can decide which of its own documents are marked superseded in a form a machine reads, which third-party listings to correct, and how the next change is published so the record moves with it. Your team makes the changes; AAAnow provides the assessment (Governance module) the evidence, the guidance and the review.
06Related use cases.
Investors acted on results the company had already superseded
The shareholder confusion this case feeds.
Brand information misunderstood by AI
The same weight-of-record mechanism, applied to the brand rather than the board.
Careers material shapes what candidates are told
The same mechanism applied to the employer.
07References.
- Outdated facts persisting in AI descriptions after the facts have changed: Lantern, 2026. www.lantern.io/blog/outdated-info-ai-searchlantern.ioWe do not enable external links - please copy and paste, you are then certain as to the link being opened
- Illustrative scenario written from the mechanism.
Next step
Talk it through, in your terms.
The briefing sets this case in your organisation’s terms, and the discussion follows your questions.
Illustrative scenario, written from the mechanism. No company is identified.