Where the book appears
Which reader questions and AI services name the book, and how consistently it recurs across repeated answers.
Services / AI discovery (GEO)
Find out which reader questions lead AI to your book
We start with one book and the questions its readers ask AI tools. We check whether the answers name it and which titles and sources they use instead, then recommend changes to the book's description and pages.
Books that explain the types of dementia in plain language
Named in 3 of 3
A short starter for the first week after diagnosis
Not named
Portable checklists for day-to-day care at home
Not named
Question lists to take to medical appointments
Not named
Paraphrased from four of the 47 questions in our study of 104 ChatGPT answers. None named the title or author.
How the analysis works
Choose the reader questions
Start with what the book helps a reader understand, do, or experience. We agree the questions to test, phrased the way a reader would ask without knowing the title.
Examine live answers and sources
Ask each question in the selected AI search services and record the books named and the sources cited. Repeating a question shows how much the answers vary.
Compare the book's supporting content
Inspect the book's description, publisher pages, and other sources the answers draw on. Separate what the evidence shows from assumptions about why a book appeared.
Deliver the findings and actions
Your team receives a report with an overview, the supporting answer and source evidence, and ranked recommendations. Proposed copy for the book remains subject to your editorial review.
What the analysis shows
A cited page and a recommended book are different outcomes
An AI answer may cite a publisher's article without naming the book it discusses. It may recommend a competing title, or name a book without giving the reader a useful route to it.
Generative engine optimization (GEO) starts with understanding those differences. We examine the questions readers ask, the books named in the answers, and the sources those answers use.
How books show up in AI searchSee how the research works
Our analysis of 27,702 ChatGPT answers found publisher websites in 75.6% of their source lists. Publisher websites were also cited in 41.5% of the 25,136 answers with traceable citations. The study examines the book pages, reading guides, and other sources that connect books with readers’ requests.
Where AI book recommendations come fromA follow-up study compared what ChatGPT wrote with the publisher pages it cited. Fewer than one cited publisher page in a hundred had an FAQ that matched the answer, so the information on a page mattered more than its question-and-answer format.
Read the publisher FAQ studyOur public study of The Dementia Caregiver examines how a book's fit with a reader's needs appears in ChatGPT answers. It shows why generic book queries and specific practical questions can produce different recommendations.
A discovery report describes observed answers and opportunities for improvement. It does not guarantee inclusion in future AI answers or establish a sales effect.
Read the reader-needs studyDiscuss AI discovery for your list
Tell Chris which books you have in mind. We can discuss the reader questions, AI services, and scope most useful to your team.
Send Chris a message