What AI tools answer when a doctor looks for a residency prep course

Two schools take up almost all the space

Estratégia MED appears in 88 of the 99 responses, and Medway in 82. Both were cited in all 25 questions, without exception: there is no question category, specialty or examining board in which either one is absent. From fourth place down, each school appears in fewer than half of the responses.

Appearing is one thing; being the first name is another

AI tools almost never answer with just one school: there are 4.1 schools per response on average. That is why where a name enters the list matters as much as whether it appears. Medway is the school that most often opens the list, with an average position of 2.0. Sanar is at the other extreme: it appears in 29 responses and enters at an average position of 4.1, meaning the reader passes three or four names before reaching it.

For a school in the middle of the table, this is where the difference lies between being considered and merely mentioned. A list has a beginning, and a name near its end competes with the names the reader has already seen.

Each tool has its own ranking

The same school changes in scale depending on where it is asked about. Grupo MedCof is the strongest case: it appears in 88% of Google’s responses and 36% of Claude’s. For anyone outside the top positions, that means there is a tool where they can compete and a tool where they cannot.

When the AI tool reads a school’s site, it almost always writes that school’s name

In 61 of the 99 responses, it is possible to see which sites the tool consulted. Putting those two things together produces a direct result: when the school’s site was consulted, its name appeared in the response 96% of the time. When it was not consulted, the name appeared in 23%.

There were 64 different sites, and the most used of all was medway.com.br , Medway’s site, in 35 of the 61. Among the ten most-used sites, 6 belong to the schools measured and 4 are outside sources, such as reddit.com, drfinancas.com and youtube.com.

Narrowing the question barely changes the answer

The 25 questions were written across five categories: general recommendation, specialty, examining board, course format, and entry examinations with prerequisites. The top three names are virtually the same in every category.

Two changes appear in the middle of the table, and they matter more to schools outside the top than to the top itself: a school that does not enter the list for the general question does not enter the list for the specific question either. There is no shortcut through a niche.

How we measure

• We asked the 25 questions in 4 tools, once each, in September 2026. That would make 100 responses; we retained 99. Perplexity produced 24 of the 25: one capture failed. Each question was asked in a clean window, in a temporary conversation with no history, from an IP in Brazil. We saved a screenshot of every one.
• No question includes a school name. All ask for a recommendation, the way a candidate asks.
• We count only when the AI tool writes the school’s name in the response. When the name appears only in the list of sites consulted, it does not enter this count: that is a different thing, measured separately in section 04.
• The table’s names came from the responses themselves. We did not start with a ready-made list, so there is no outside school that "should" be there: if it was never cited, it does not appear.
• Two companies with two active names count as one because AI tools themselves write them together: MedGrupo and MedCurso on the same line, and CRMedway within Medway.

What this study does not say

• Nothing about teaching quality or admission results. What is measured is what AI tools answer, not who teaches better or who gets more students admitted.
• Nothing about how much each thing is searched for. The 25 questions represent how a candidate asks, not a sample weighted by search volume.
• Nothing about what the AI tool answers for you. Responses vary with the previous conversation, region and day. This is a September 2026 snapshot from a clean window.
• Nothing about how to change this result. This material measures; it is not an action plan.

All 25 questions, in full

Original Portuguese research evidence, preserved as recorded.

1. 01melhor curso para passar na residencia medica
2. 02melhores cursinhos para residencia medica
3. 03qual cursinho de residencia medica mais aprova
4. 04onde estudar para a prova de residencia medica em 2026
5. 05qual o melhor curso online para residencia medica
6. 06me indique um cursinho para residencia medica
7. 07qual cursinho de residencia medica vale a pena contratar
8. 08melhor curso para passar em residencia medica de ortopedia
9. 09melhor cursinho para quem quer dermatologia na residencia
10. 10qual curso indicar para quem vai prestar oftalmologia na residencia
11. 11melhor preparatorio para residencia medica de pediatria
12. 12melhor curso preparatorio para o enare
13. 13qual cursinho prepara melhor para a prova de residencia da usp
14. 14melhor curso para a residencia medica do sus-sp
15. 15qual curso preparatorio indicar para o amrigs
16. 16melhor cursinho para a residencia medica do einstein
17. 17melhor curso extensivo para residencia medica
18. 18melhor intensivo de residencia medica para quem esta no ultimo ano da faculdade
19. 19qual o melhor banco de questoes para residencia medica
20. 20melhor curso presencial de residencia medica em sao paulo
21. 21qual plataforma de questoes comentadas para residencia medica vale a pena assinar
22. 22melhor curso para a prova de r+ de cardiologia
23. 23melhor preparatorio para r+ de endocrinologia
24. 24qual cursinho indicar para quem vai prestar r+ depois da clinica medica
25. 25qual curso prepara para a prova pratica e a segunda fase da residencia medica