A single screenshot can show that an AI-assisted search service found and used information about a funeral home on a particular day. It cannot prove a permanent recommendation, full question coverage or a commercial result. Treat it as one observation, not the strategy or the score.
Results need context because the wording of the question, the searcher’s location, the date and the service being used may all affect what appears. A funeral home can be named for one question and absent from another closely related one. That is why a useful review starts with a planned set of questions rather than a search chosen after a favourable answer has appeared.
The agreed scope should name the search experiences being checked. For Funeral Marketing Works that means ordinary Google Search and Maps alongside selected checks in ChatGPT, Gemini, Perplexity and Google’s AI experiences. Interfaces and availability change, so every observation should retain its platform and date instead of being presented as a permanent platform-wide result.
The terminology is less complicated than it sounds. SEO, or search engine optimisation, improves how a business can be discovered in conventional search. AEO, or answer engine optimisation, helps clear information answer the questions people ask. GEO, or generative engine optimisation, considers how AI-generated search experiences find and combine supporting information. The labels overlap; for a funeral home, the sensible work is one connected programme of accurate facts, helpful answers and credible evidence.
Build the question set around genuine family needs. It can cover immediate help, local availability, a particular service, practical arrangements, price information and reasons to trust a provider. Record the intended town or area with each local question. The aim is not to create hundreds of artificial prompts, but to keep a representative set that can be repeated on comparable terms.
The practical work sits underneath the answer. The funeral home’s own website should state its locations, services, people and contact details clearly. Its Google Business Profile and other public listings should agree with those facts. Genuine reviews, responsible replies, useful service and area pages, local citations and reputable independent mentions can all help a person or a system understand what the business is and where it operates.
A baseline can be simple and still be disciplined. For each question, record the exact wording, platform, date, location context, businesses named, order where that is meaningful, sources shown and any inaccurate or missing facts. Save the evidence with a short note explaining what can reasonably be learned from it. Do not turn an illustrative dashboard or a selected screenshot into an unsupported visibility percentage.
Repeat the same core checks at an agreed interval and compare the pattern with ordinary Search and Maps visibility, website visits and worthwhile enquiries. If mentions increase but calls and forms do not, the next question is whether the search intent, website journey or measurement needs attention. AI visibility should not be reported in isolation from the commercial journey it is meant to support.
Prioritise corrections that help families as well as machines. Fix a wrong address before writing more content. Make an important service easy to find before producing a speculative article. Strengthen an unclear contact route before celebrating an additional citation. Each action should have an owner and a reason, with a distinction between an immediate correction and a longer-term opportunity.
Avoid shortcuts that damage trust: manufactured reviews, invented local offices, copied town pages, paid mentions presented as independent evidence or synthetic staff presented as real people. Structured data can reinforce information that visitors can already see, but it cannot make an unsupported claim true or guarantee that a service will quote the business.
The honest objective is wider, more dependable discoverability across Search, Maps and the agreed AI experiences—not control of a generated reply. A good review shows what was checked, what the evidence supports, what remains uncertain and which three improvements deserve attention first.