The map pack and the AI answer are different games — play both

Two doors, one house

Two families in the same county open their phones on the same evening. The first types “funeral homes near me” and gets the familiar map — three pins, three names, a row of stars. The second asks an assistant “who should I call after a death at home in [your town],” and gets a short paragraph naming two funeral homes, a sentence on each. Both families are doing local search. They are not looking at the same thing, and the two things are not won the same way.

The proportions are worth stating before the tactics. For plain local searches — the “funeral homes near me” kind — Whitespark found Google’s AI answers appearing only about 15% of the time, while the classic map pack showed in 93%. The map is still the bigger door, by a wide margin. The paragraph is the one growing — and the one almost no funeral home is tending.

The map pack runs on proximity, categories, and profile signals — the classic local playbook. How close your building is to the searcher, which categories your Google profile claims, how complete and active that profile is. The AI answer runs on language: what your pages say in plain sentences, what reviewers describe in their own words, how consistently the rest of the web agrees on your facts. One is a question of place and profile; the other is a question of prose and consensus.

This matters because the two can disagree. A funeral home can own the map — closest pin, most reviews, every profile field filled — and never appear in the paragraph, because a complete profile can sit above a public record that says almost nothing. Another home, a little farther out, can be named in the paragraph again and again because its pages actually answer what families ask. Winning one game tells you little about the other.

Here is the useful way to hold it. Your funeral home has one set of facts — its hours, its address, its reviews, its services — and each game reads that same set differently. The work ahead is not two marketing campaigns. It is understanding what each door asks of the same house.

What serves both

Start where the games overlap, because that work pays twice. Four things feed the map’s algorithm and the model’s reading at once, and none of them needs an agency.

One canonical name, address, and phone, identical everywhere — your site, your Google profile, the directories. The map cross-checks these to trust where you are; the model cross-checks them to trust that you exist and are where you say. A single digit wrong in one directory is a small cost in both games at the same time.

A Google Business Profile that is current and answered — hours right, categories right, questions replied to. Completeness is a direct map signal, and the same profile is the grounding data Google’s own AI answers read when they describe you. (ChatGPT, notably, never reads it — it reads the open web — which is exactly why the profile alone doesn’t win the paragraph.)

Reviews with sentences in them. This is the clearest case of one action doing two jobs. A review that says “they picked up my mother at two in the morning and had everything arranged by Thursday” adds to your count and your recency, which the map weighs — and it hands the model a specific line to quote about what you actually do, which “wonderful people, five stars” never will. You cannot write your reviews, but you can invite the kind that describe what happened, and those serve both doors.

Service pages that say plainly what you do and where. “We serve [your county] from our chapel on Main Street; we answer the phone around the clock and handle burial, cremation, and pre-planning” tells the map which categories are true and hands the model sentences to lift. Same paragraph, two readers.

Do this shared work first. For most homes it closes most of both gaps.

What moves the pin but not the paragraph

Some of the map’s strongest signals do nothing for the AI answer, and it helps to know which — so you neither over-invest in them nor mistake a map win for a paragraph win.

Proximity is the largest, and the least fair. The map heavily rewards the funeral home nearest the person searching, and you cannot move your building. A home three blocks from the county’s densest neighborhood will out-pin a better home on the edge of town, for reasons that have nothing to do with care. Categories come next — the service types your profile claims decide which searches you enter at all — and profile completeness, photos, posts, and a steady trickle of fresh reviews each nudge the pin.

The quiet point is that almost none of that reaches the paragraph. A written answer from ChatGPT or Perplexity is not ranked by how near you are to the reader; it is composed from what the web says about you. So the map rewards a good deal you cannot change, or can only tend, while the paragraph rewards the thing you can most control — the words on your own pages. For a funeral home that will never be the closest pin, that is not bad news. It is the opening.

What serves the paragraph alone

The AI answer wants one thing the map never asked for: prose it can lift. The map is content with fields and pins; the paragraph needs sentences.

So the pages the map barely notices are the ones the paragraph feeds on — your FAQ page, your plain-language explainers of what happens after a death at home or what cremation costs here, your steady obituary page proving the home is active this week. A machine composing one paragraph reads these, finds the sentence that answers the family’s question, and quotes it — with your name, if the sentence named you.

Here is the move that organizes all of it. Write the two sentences you hope the assistant will say about your home, then publish the evidence for them. If you want it to say “they answer the phone at any hour and explain costs plainly before you decide,” then your hours, your after-hours promise, and your price explanation had better exist in plain words on a page of your own. The paragraph cannot lift what you never wrote down.

Measure each separately

Because they are two games, you check them two ways, once a quarter, with a notepad.

For the map: search your main terms — “funeral home near me,” “cremation in [your town]” — from a phone in your service area, and note where your pin falls and who sits above you. For the paragraph: ask ChatGPT, Gemini, and Perplexity the questions families now put to them directly — “who do I call after a death at home in [your town],” “best funeral home in [your county]” — and note whether you are named, and whether every fact about you is right.

Two lists, two sets of gaps — and, most quarters, nearly one to-do list, because the shared work moves both. Write down where you stand in each. A funeral home that watches both doors honestly, four times a year, is already ahead of nearly every competitor it has — most of whom are watching neither.