01What happens.
AI answers from what it can reach and cites what it can identify. A document without a title or an author is hard to cite, and a site that turns AI away cannot be reached. Where the company's own material is absent, whatever else exists takes its place, and a version built to be cited is cited.
02Why it matters.
If you are the finance director on results day, the summary an investor reads was assembled by a machine from whatever it could find under your company's name. If your release was unreachable and unlabelled, it found something else.
A version of your results written by someone with an interest in the share price, structured better than yours, is a version AI has no reason to distrust.
The market reads someone else's account of your numbers, and the correction has to begin by proving what you published.
03The case.
This illustrative scenario draws on 2 documented conditions and no company is identified.
The first condition is the digital doorway. AAAnow's outside-in assessments of organisations that appeared fully operational to people found AI routes turned away or left without a reply. In 5 anonymised assessments, between 4 and 14 of the 15 tested AI-related routes were turned away while people were served. A company in that position has restricted the opportunity for its current, controlled information to shape the answer, and has not restricted the answer.
The second condition is the fake built to be cited. Aurascape documented the first campaign in which structured fakes were planted for AI systems to cite as official, an actor publishing documents built to a machine's preferences so that AI would treat them as the source.
Put the 2 together on results day.
The company posts its results as PDFs with no document title, no author and no metadata, on a site that turns the AI routes away. A hostile party, positioned on the share price, publishes a results summary elsewhere: titled, dated, structured, with the figures altered where it matters.
AI, asked how the company did, reaches the hostile version, cannot reach the company's, and answers from the one it has. The finance director reads an AI summary of the company's results that quotes figures the company never released, in a form that looks like the company's own. Investors have read the same summary, and some have traded on it.
The company's release, correct in each figure, is on a site the machine was told to leave.
Nothing in the company's own checks would have shown this, because the site worked for people, the PDFs opened and the release was accurate. What the company could not see was that, to the systems now answering investors' questions, it had become unreachable, and another source had taken its place.
04Exposure.
A version of the company's results in circulation that the company did not issue, attributed to the company, is a direct market-integrity exposure. The board answers for the disclosure record and now has to establish, in public, which version is the company's. Whether any obligation is engaged on the facts is for the company's advisers; the position the board is in is not in doubt.
The cost begins with the correction and does not end there. Investors who traded on the hostile version, an adviser who repeated it, and a regulator asking why the company's own release could not be found, are each a cost the blocking decision created. In industry benchmarking, 57% name AI errors and misinformation as the top concern threatening AI adoption (Gallagher 2026, respondents), and the covers a board relies on are being rewritten with that in mind. No outcome is automatic.
05What AI Readiness changes.
What the company cannot see from the inside is that its digital doorway is closed to the systems answering investors' questions, and that its results documents cannot be identified once they are f ound.
AAAnow tests the routes AI uses from the outside, 15 of them, and evidences whether each is served, challenged or turned away. It reads the company's documents as AI reads them and evidences what can and cannot be cited. With that evidence the board can decide which restrictions it wants, record each one and publish its results in a form the machine identifies as the company's own.
Your team makes the changes; AAAnow provides the assessment, the evidence, the guidance and the review.
06Related use cases.
£350,000 to be seen by AI routes the estate was turning away
Spend on visibility while the doorway was shut.
Investors acted on results the company had already superseded
The disclosure record read wrongly from the company's own material.
Announcements buried in the page furniture
The announcement that was published and never heard, on a page that carried too much else.
07References.
- AAAnow outside-in digital doorway assessments, 5 anonymised examples, Jul. and Aug. 2026, published in Lawrence Shaw, Digital Suicide, LinkedIn, 3 Aug. 2026.
- Aurascape, the first documented campaign planting structured fakes for AI systems to cite as official.
- Gallagher, AI adoption and risk benchmarking 2026. www.ajg.com/news-and-insights/features/ai-adoption-and-risk-benchmarking-2026ajg.comWe do not enable external links - please copy and paste, you are then certain as to the link being opened
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; the mechanism is documented. No company is identified.