01What happens.
A PDF carries more authority in AI results than the page that linked to it. The page said which year the report belonged to; the PDF says nothing about its own age that a machine reads. Remove the page and the PDF keeps answering, with the authority of the format and without the context.
02Why it matters.
If you run investor relations or the company website, the archive pages you removed in the redesign were the only reliable context telling a machine which report belonged to which year. Your document store holds each year's report for a reason, and each of them is now a candidate answer to the question of how the company is doing.
PDFs are cited by AI at 1.4 to 1.8 times the rate of comparable web pages, on AAAnow's own reading of more than 4.5 million PDF pages, and other research puts the figure at up to 2.5 times. The format is trusted more than the page that framed it.
03The case.
This illustrative scenario draws on patterns observed across listed company estates; no company is identified.
A listed company rebuilds its investor site. The old site had an archive: a page per year, each with a short summary and a link to that year's report. The new site is cleaner, and the archive pages go, and the current report is presented on the results page with a link to the document store for anything older.
The document store keeps each year's report, because the company is required to keep them available. Each report is a PDF, produced by that year's design agency, with no date in its file properties and, in the older ones, no year on the cover that a machine reads. The store lists them by filename, and the filenames follow whatever convention each year's agency used.
An investor asks AI what the company's revenue is and what its strategy is. AI reaches the document store, finds a set of annual reports, and weights them as a machine weights PDFs: highly, and without a date to separate them. The answer draws on the report from 5 years ago, because it is the longest, the most linked from elsewhere on the web, and the most cited over time.
The strategy AI describes was replaced 2 reviews ago, and the revenue is from a business twice the size the company is now.
The investor relations manager who ran the redesign checks the new site and finds it correct. The current report is where it should be. The archive is in the document store where the auditors wanted it.
Nothing on the site now says reliably that the old reports are old. A filename is not a date a machine trusts, and until the redesign the archive pages had supplied the context; nobody thought of those pages as part of the disclosure.
Each old report now answers as current to a machine that reads the format's authority and none of its history. The more thorough the archive, the more old positions are available to be quoted, and the company that kept its records most carefully is the one most exposed to them.
04Exposure.
The company is held to what it published, and each report in the store was published. A 5-year-old strategy or revenue figure quoted as current, under the company's name, is a misstatement in circulation that the company issued, in a form the company still hosts. The board's accountability for the disclosure record includes the record's history where that history is served as the present.
The cost is correction, one investor at a time, of a position the company held and moved on from. The heavier cost is confidence in the company's current disclosure, which is discounted when the market's picture of the company is 5 years out of date and the company appears not to have noticed.
05What AI Readiness changes.
What the company cannot see from the inside is that its archive is answering as its present. AAAnow reads the company's digital landscape as AI reads it and evidences which documents AI is drawing on, how each is weighted, and which of them carry no date a machine reads. With that evidence the company can decide which reports stand as the current record, and mark the rest with the year, the framing and the superseded marking a machine reads, before the next investor asks.
Your team makes the changes; AAAnow provides the assessment, 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.
Decades of documents that outlive the context that framed them.
Duplicates and untracked copies
The same report survives in copies the company cannot see.
07References.
- PDF citation authority 1.4 to 1.8 times comparable web pages: AAAnow analysis across more than 4.5 million PDF pages. Up to 2.5 times: other research, cited as such.
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, drawn from observed patterns. No company is identified.