Which service should I prioritize watching?

Domestically, the service that reaches the most users is Google (AI Overview and AI Mode). Perplexity is where whether your sources and author information are clearly stated tends to be questioned, and ChatGPT is observed to be strong for consultations using proper nouns. Rather than narrowing it down to one, watching across several is the realistic approach.

Do clicks from AI search really decrease?

The behavior varies greatly by the type of query. For fact-checking and definition-type queries, a decline in clicks is observed, but for service-comparison and booking-path queries, a certain amount of traffic to the cited sources remains. For clinics, the latter tends to be the main case.

Is observation better done in-house or outsourced?

To preserve the reproducibility of observation, at first we recommend bringing in outside expertise for designing the query set and standardizing the observation procedure. Once the procedure can be standardized, ongoing observation thereafter can be handled in-house.

The Differences in AI Search That Medical Institutions Should Watch

When a patient consults an AI about a medical institution, what kind of answer comes back changes depending on the service used. Even for the same question, the summary shown on Google's search screen, a conversation with ChatGPT or Claude, and Perplexity's answer all differ in both the information referenced and how it is presented. Here, we organize the differences among the main services that a medical institution should know when checking how its own clinic appears.

The way it appears differs from service to service

AI search and AI assistants share the common point of "summarizing multiple pieces of information together and, as needed, indicating the referenced sources." However, what kind of information they reference easily, how they show sources, and what kind of questions they excel at differ from service to service.

For that reason, you cannot judge that "because it appeared on one of them, we're fine" or "because it didn't appear, we failed." The realistic approach is to look across multiple services at how your own clinic is handled.

Google: The difference between AI Overview and AI Mode

Google has two AI features with different characters. At present, it is easier to understand if you grasp them as follows.

AI Overview

This is an AI summary displayed within ordinary Google search results.

When a user searches, it may appear within the search results screen
It may organize information accompanied by links to the referenced sources
It is easier to understand if you think of it as "an extension of ordinary search"
In medical searches, behavior that keeps the display more restrained is also observed

AI Mode

This is a more conversational, exploratory search experience.

It suits complex questions, or a usage where you investigate while stacking up follow-up questions
Within the back-and-forth with the AI, you dig deeper into the information
Rather than an ordinary search results screen, it is closer to an experience of "investigating while consulting the AI"

Since the specifications of both can change, for your own clinic it is most reliable to actually search and confirm while recording how it is displayed.

How to view ChatGPT, Claude, and Perplexity

ChatGPT

This is a service where search, answering, and deeper investigation continue along the flow of conversation. It is strong for consultations using brand names, service names, and proper nouns, and it tends to return detailed answers for questions that require explanation.

Claude

This is a conversational AI assistant. Rather than a search engine, you view it as a target that organizes information and answers in response to a consultation. It becomes a target for confirming how medical-institution information is summarized and compared within a conversation.

Perplexity

This is a service where, as a basic rule, answers are accompanied by the URLs of the referenced sources. Rather than being the center of general use, it is worth watching for being a service where it is easy to verify "which information source was cited." It has been pointed out that it tends to reference pages where sources and author information are clearly stated.

What to look at on a medical-institution site

When confirming how your own clinic is handled, the following elements are thought to be referenced easily.

Element to watch	Google AI Overview	Perplexity	ChatGPT
Structured data (Schema.org)	◎	○	○
Clear statement of author information	○	◎	○
Listing of sources alongside	○	◎	○
Consistency of names and notation	◎	◎	◎
Recency of the update date	◎	◎	○
Links to public institutions	◎	◎	○

Among the articles cited by Perplexity, there was a tendency for a high proportion to clearly state author information. This suggests the possibility that, when an AI references an article, it handles pages where it is clear who wrote it more easily. On your own clinic's site too, clearly showing the operator's and author's information is a point you will want to confirm.

Perspectives when comparing

When comparing multiple services, recording from the following perspectives makes it easier to grasp the trends.

Whether your clinic's name appears as a candidate
For what kinds of questions it appears / does not appear
Which other medical institutions appear alongside it
Where the referenced information comes from (official site / reviews / map information, etc.)
Whether there are errors in the content of the answer

Summary

How AI search and AI answers appear differs from service to service
Google has a summary within the search results (AI Overview) and a consultation-type AI Mode
With ChatGPT, Claude, and Perplexity too, the way information is handled differs
Rather than judging by a single service, confirm while recording across several