# How Brazil creates courses with AI: 6,821 people

> Brazil is the largest country in Coursebox's course creation data (18%). What Portuguese speakers build with AI and how they brief it.

**Author:** Parshad Desai  
**Published:** 2026-11-09  
**Categories:** Course Creators, Training Organisations

---

Brazil is the largest country in Coursebox's course creation data. More people create courses with AI from Brazil than from India or the United States.

I looked at the anonymised first prompts of 37,284 people who created a course with AI on Coursebox between May and November 2026. 92% of the Portuguese-language prompts came from Brazil. **Portuguese speakers build courses about earning income and personal development at about twice the rate of English speakers, and 49% of them write their first prompt on a phone.**

> From the Coursebox Research Report: [The State of AI Course Creation 2026](https://www.coursebox.ai/research/state-of-ai-course-creation)

## How big is course creation with AI in Brazil?

18% of the people who created a course with AI on Coursebox are in Brazil, the largest single country, according to Coursebox's data on 37,284 people who created a course with AI (May to November 2026). India follows with 10% and the United States with 9%. Portuguese is the language of 19% of first prompts, and Brazilian Portuguese alone accounts for 13% of all courses generated.

_Share of people creating courses with AI, by country_

| Category | Value |
| --- | --- |
| Brazil | 18.3% |
| India | 9.7% |
| United States | 9.3% |
| Philippines | 7.9% |
| Colombia | 5.3% |

Base: first prompts from 37,284 people, May to Nov 2026

Those courses are written in Portuguese from the start. 97% of people in Brazil write their first prompt in Portuguese, so the course is built in the learner's language from the first draft.

## What do Portuguese speakers create with AI?

Portuguese speakers create job skills courses as often as English speakers, and courses about earning income and personal development about twice as often. Earning income accounts for 14% of Portuguese learning prompts against 6% in English, and personal development for 15% against 8%. Academic subjects are less common, at 7% against 17%.

_What the course is for, by prompt language_

| Category | Portuguese | English |
| --- | --- | --- |
| Job and technical skills | 28% | 26% |
| Personal development | 15% | 8% |
| Earning income | 14% | 6% |
| Academic subjects | 7% | 17% |
| Compliance and safety | 6% | 6% |

Base: 1,464 Portuguese and 3,014 English learning-related first prompts, from 10,000 classified, May to Nov 2026

The sectors follow the same pattern. Business and entrepreneurship is the largest identifiable sector in Portuguese prompts (13%), followed by technology (10%) and beauty, wellness and fitness (8%).

## Who creates courses in Portuguese?

Independent course creators make up 17% of Portuguese learning prompts, more than twice their 8% share in English. Teachers are less common (3% against 8%), and 45% of Portuguese prompts are for the general public, against 22% in English.

_Share of learning prompts by creator type_

| Category | Portuguese | English |
| --- | --- | --- |
| Course creators and coaches | 17% | 8% |
| Corporate L&D | 6% | 8% |
| Teachers and lecturers | 3% | 8% |
| Training providers | 3% | 5% |

Base: 1,464 Portuguese and 3,014 English learning-related first prompts, from 10,000 classified, May to Nov 2026. Creator type is identifiable in about a third of prompts

Among Portuguese-speaking creators, 47% build courses about earning income and 22% about personal development. Typical examples include a beginner's course on selling through an online marketplace, a course on starting a small business from home and a course on managing anxiety.

## How do Portuguese speakers brief the AI?

Portuguese speakers write shorter first prompts and more of them write on a phone. The median Portuguese first prompt is 16 words against 27 in English, 49% are written on a mobile device against 40%, and 12% attach a file against 19%.

_How the first prompt is written_

| Category | Portuguese | English |
| --- | --- | --- |
| Written on a phone | 49% | 40% |
| File attached | 12% | 19% |
| Covers none of 5 design decisions | 49% | 43% |

Base: first prompts from 7,213 Portuguese and 17,099 English speakers; design decisions from 1,074 and 2,269 classified learning prompts, May to Nov 2026

A short brief on a phone still produces a full course. More of the course design simply happens in the follow-up conversation with the AI.

## What should you do if you create courses for Brazil?

Whether you sell courses in Brazil or train a Portuguese-speaking team, these steps fit how the market already works:

- **Build in Portuguese from the start. **Write the brief in Portuguese so the course, examples and quiz come out in the learner's language rather than as a translation.
- **Design for the phone. **Half of Portuguese speakers start on a phone, and many of their learners will study on one. Keep pages short and quizzes tap-friendly.
- **Lead with the outcome. **Many Portuguese-language courses are built to be sold on a clear result: a new skill, a first sale, a healthier routine. State that result in the brief.
- **Add your own material. **Only 12% attach a file. Your method, examples and stories are what separate your course from the next one on the same topic.

## How I measured this

The data is anonymised and aggregated. No names, email addresses or company details were used in the analysis, no prompt text is published, and any group smaller than 30 is left out. It covers 137,126 course-creation prompts from 37,284 people on Coursebox between 14 May and 9 November 2026, with bots, staff and test accounts removed. Country comes from the visitor's IP address and language from the prompt text and browser language. Categories come from an AI classification of a random sample of 10,000 first prompts written in people's own words; a manual check of 40 classifications from each of the two AI models found about 92% correct, and 304 prompts labelled by both were checked by hand to calibrate the design-decision figures. The examples above are paraphrased and generic. The full methodology is in the report.