01What happens.
When municipalities consolidate, their websites, microsites and documents come with them. AI reads whatever is reachable and weights most heavily the material that is well structured and widely cited, whether or not it is current. A well-built old site outranks a weakly built new one.
02Why it matters.
Your residents, and the agents and chatbots acting for them, act on the answer AI returns. When that answer comes from a site you do not know you have, the denials, the appeals and the lawsuits come back to you. Your own new systems repeat the wrong answer, because they read the same estate.
The estate you inherited is part of the position you are judged on from the first day of the new municipality.
03The case.
Several municipalities were consolidated into one. The new city took on the digital landscape of each of them: websites, microsites, departmental sites and the documents they carried. No single map of that landscape existed.
Among the inherited properties was a planning and zoning microsite built some years earlier by a department of one of the original municipalities. Its documentation was complete and in order, it was widely linked and cited, and conventional search had served it for years. AI systems treated it as the master source for planning in the area, and the city had forgotten it was there.
The new city’s official planning and permitting portal carried none of that standing. Some documents had version numbers, some were dated and some carried neither, and against the old microsite the official portal carried less weight with AI.
Residents asking AI about planning received the old municipality’s rules and procedures as if they were current, and the new city’s own chatbots served the same information. Permit applications were prepared to the wrong requirements and denied. Appeals followed, some cases went to litigation, and residents’ experience of the new city was poor from the start.
In 2 instances the effect was compounded by access: on one property AI was fully blocked and on another AI search was blocked, so those routes fell back on the older material open to them. Blocking was a contributory factor and the landscape was the cause.
Multiple versions of the same content sat across inherited properties, no map of what the city had online existed, and the new portal was published without the version and date signals a machine reads. The digital team had not been asked to map the estate, and no tool showed them what AI was reading.
04Exposure.
A municipality answers for the planning information published under its name, including on properties it has forgotten. Denials made on wrong advice are the city’s denials, appeals and litigation are handled at the city’s cost, and where the city’s own chatbot repeats the wrong answer, the city has published it a second time. New York City’s business chatbot showed the same failure on the public record: it misstated city policies and advised businesses to break the law, staff asking it the same housing-voucher question received different answers at different times, and it was taken down (Associated Press via NBC New York, 2024; The Markup and StateScoop, Jan. 2026).
Each wrong denial is reworked: the appeal, the staff time, the correspondence and the legal counsel where a case proceeds. The wider cost is the confidence of residents in a new city at the moment it needs to establish itself, spent on a website the city did not know it had. Across risk profiling of more than 119 million websites, 41% were unknown to the organisation’s own digital teams (P&C / Sitemorse risk profiling, 2017 to 2023).
05What AI Readiness changes.
What the city could not see from the inside was which of its inherited properties AI was treating as the master. AAAnow reads a municipality’s digital landscape from the outside, as AI reads it, and evidences the properties, documents and versions AI is drawing on, including those absent from the city’s own records. With that evidence the city can decide which planning position stands, retire the properties that contradict it, and mark its current documents with the version and date signals a machine reads.
Your team makes the changes; AAAnow provides the assessment, the evidence, the guidance and the review.
06Related use cases.
The chatbot was replaced twice before anyone looked at what it was reading
The same failure in a UK authority, and the cost of chasing the wrong cause.
The estate the digital team had warned about for years
Mapping the footprint as the first step, and the return when it is done.
New York City’s business chatbot told companies to break the law
The same failure on the public record, with the chatbot withdrawn.
07References.
- Associated Press, reported by NBC New York, 2024. www.nbcnewyork.com/news/local/nycs-ai-chatbot-was-caught-telling-businesses-to-break-the-law-the-city-isnt-taking-it-down/5287713nbcnewyork.comWe do not enable external links - please copy and paste, you are then certain as to the link being opened
- The Markup and StateScoop, Jan. 2026: the same housing-voucher question answered differently to 10 city staff.
- 41% of websites unknown to the organisation’s own digital teams: P&C / Sitemorse risk profiling, 2017 to 2023, covering more than 119 million websites.
Next step
Talk it through, in your terms.
The briefing sets this case in your organisation’s terms, and the discussion follows your questions.
Observed, anonymised. No organisation is identified.