01What happens.
A discontinuation and a disposal are announced once. The product and the division remain described across product pages, catalogues, reviews, press coverage, annual reports and directories. Historic records remain accurate for their period, but without a clear machine-readable relationship to the announcement, AI can present them as current.
02Why it matters.
The product director closed the line and the corporate team completed the disposal, both properly. The wider record still contains far more descriptions of the product and division while active than statements of their ending, and AI can present that historical record as current.
Outdated facts persist in AI descriptions after the facts have changed, and discontinued products are among the facts documented as persisting. For a consumer brand the cost lands in customer contact and retailer relationships; for the listed parent it lands in the shareholder confusion of use case 32.
03The case.
No company is identified in this illustrative scenario. It applies the mechanism documented by Lantern: discontinued products persisting in AI answers after the underlying facts have changed.
Line discontinued, product pages are taken down, the retailers are notified, stock is run down, and the discontinuation is announced in the trade press. A year later the group sells the division that made the line, and the disposal is announced to the market.
The product survives across the wider record: in 2 years of catalogues held by retailers, in review and comparison sites, in the brand's own press archive, and in a support microsite that was never taken down. So does the division, in years of annual reports, in the group's investor presentations, in press coverage and in the same third-party directories.
A customer asks AI whether the product is available and is told that it is, with a description drawn from the catalogue copy and a link to a retailer listing that no longer resolves. The customer contacts the retailer, who explains, and then the brand, who explains again, and the brand's customer team logs another enquiry for a product it stopped making a year ago. A second customer asks which company makes it and is told the group does, through the division it sold.
An investor asks AI what the group's divisions are and receives the sold division among them, with its revenue from the last annual report in which it appeared, presented as part of the group's current figures.
The brand and the group had each completed the formal process and had records to show it. The process ended at the announcement, but it did not establish how current pages and machine-readable records should distinguish the new position from accurate historical material.
04Exposure.
A brand has to correct outdated product information still circulating under its name. In this scenario, retailer catalogue copy says the product remains on sale, while the brand's support microsite and press archive supply product information that AI also treats as current. For the listed parent, the 6 annual reports are accurate historical records; the error arises when AI presents the sold division and its past revenue as part of the current group.
Customer contact and retailer correspondence are the brand's cost; investor questions and correction of the group's current description are the parent's. Both follow from AI treating historical and superseded material as current after the announcement, rather than from accurate historical documents remaining unchanged.
05What AI Readiness changes.
The announcement was made and nobody listed the record it needed to change. Read from outside, as AI reads it, the record of the product and the division shows which pages, catalogues, reports and directories AI is drawing on. It shows which of them the brand and the group control, and which belong to retailers and third parties. AAAnow evidences that reading and what AI answers.
With that evidence the brand retires the obsolete support microsite, marks relevant archive material as historical, and takes the evidence to retailers and directories. The group preserves its historical disclosure record and publishes the disposal in a machine-readable form that distinguishes the current structure from past reports.
The company makes the changes, and 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 that the division side of this case feeds.
Customers were shown a brand the company had retired, from guidelines it had replaced
The weight of the record outvoting the current page.
The product announcement was published and missed, on a page that carried too much else
The announcement that did not reach the machines.
07References.
- Lantern, 2026: outdated facts, including discontinued products, persisting in AI descriptions after the facts have changed. https://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
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.