CASE STUDY · 2026 · PROGRAMA GANHE O MUNDO, PERNAMBUCO
How Pernambuco Used AI to Prepare Students for Global Exchanges
Starting in September 2025, Pernambuco's Ganhe o Mundo program introduced personalized English practice through WhatsApp to prepare students for international exchanges. Across six learning experiences, 537 students practiced real-world communication scenarios, generating more than 40,000 messages and interactions over four to five months.
- 537
- STUDENTS REGISTERED
- 94%
- REPORTED ENGAGEMENT RATE
- 40K+
- MESSAGES & INTERACTIONS
- 73%
- DEEP ENGAGEMENT
- 4–5 months
- ACCESS PERIOD
Public-school students preparing for international study
Strong participation across the registered student population
Generated across six targeted learning experiences
Students completing more than 20 interactions during the final two weeks measured
Beginning in September 2025
Context
About the program
Programa Ganhe o Mundo is an international-education initiative created by the Government of Pernambuco to give public-school students access to language learning and study abroad. Established by state law in 2011, the program was designed to make international mobility possible for students who might otherwise have few chances to live and study in another country. Selected students continue their studies abroad, experience another culture, become more independent, and use English or Spanish in everyday life.
Preparing students across Pernambuco
Before the 2024 selection process, more than 20,000 students took preparatory English and Spanish classes at 226 learning centers across Pernambuco's 16 regional education districts. These included centers in Indigenous territories and virtual English and Spanish classrooms designed to reach students across the state.
A statewide selection process
A total of 9,727 students took the 2024 selection exams, an objective language assessment covering reading, comprehension, interpretation and vocabulary in English or Spanish. Eligibility also took into account school attendance, academic performance in Portuguese and mathematics, and age at the time of travel. In the end, 918 students were approved for exchanges in Canada, the United States and Chile. For the first time, each of Pernambuco's 184 municipalities, plus Fernando de Noronha, had at least one student among those selected.
- 20,000+
- Students in preparatory classes
- 226
- Learning centers
- 9,727
- Students in the 2024 selection exams
- 918
- Students approved for exchanges
Challenge
Moving from grammar to real-world communication
English was already part of students' education. But much of what they had learned emphasized grammar rules, written exercises, vocabulary lists and familiar textbook topics. Many students had studied English for years and could recognize vocabulary, understand grammar rules or answer written questions. Far fewer had regular chances to hold conversations, express themselves spontaneously, understand different speakers or ask for help in English. That gap matters once students arrive in homes, schools and communities abroad, where English is the language they need to get through everyday life, not just a school subject.
Three structural barriers
"Students did not necessarily reject English itself. Many were reacting to a learning experience dominated by rules rather than communication."
Solution
Real-world English practice through WhatsApp
Pernambuco introduced Jinso because it could provide individualized English practice across the state without requiring every student to attend extra live classes or use a computer lab. Jinso met students in a place already central to their daily communication: WhatsApp. The AI tutor adapts to each learner's proficiency, supports written and spoken practice, and guides students through scenarios linked to their upcoming time abroad. Educators and administrators can also see participation, recurring language difficulties and areas where extra academic support may help.
Why Jinso?
- WhatsApp-native delivery: Students reached the learning experiences through a familiar channel, with no separate app to download or learn.
- Real-world scenarios: Introducing themselves, communicating with a host family, describing routines and preferences, asking questions, making polite requests, taking part in school activities, and handling unfamiliar situations.
- Written and spoken practice: Students practiced forming complete responses, recording audio, understanding conversational prompts and expressing their ideas, not just multiple-choice or grammar activities.
- Personalized difficulty: Jinso adapted its language, prompts, explanations and support to each learner's level. Students who needed more help got simpler prompts and extra clarification, while advanced learners worked on more complex scenarios.
- Anytime access: Students could use Jinso at home, in the evening, on weekends or during school holidays.
Student user journey
- Access: Students get access directly through WhatsApp, which reduces onboarding friction.
- Practice: Learners complete personalized scenarios through written messages and audio responses.
- Feedback: Jinso provides immediate corrections, explanations and chances to try again.
- Develop: The experience adapts to each student, and educators can see engagement and recurring areas of difficulty.
"WhatsApp made it possible to reach students quickly through a channel they already understood, using only a mobile phone and basic connectivity."
Implementation
What made this work: practice connected to the program
Jinso wasn't introduced as a generic language-learning app. The experience was designed around the specific communication needs of Ganhe o Mundo students and the situations they would face abroad. Four implementation principles helped turn access into real participation:
- Program-aligned learning experiences. Jinso delivered six targeted learning experiences tied to students' preparation needs. Instead of drilling disconnected vocabulary, students worked through realistic situations. The Host Family lesson was the most-practiced experience, both by number of learners and by total interactions.
- Continuous program intelligence. The platform surfaced aggregate patterns in sentence construction, message length, vocabulary development, recurring grammar difficulties, practical communication, engagement levels, and areas where extra instructional support could help.
- Practice without public pressure. Students could repeat activities privately and get feedback without being evaluated in front of classmates.
- Rapid, low-friction activation. Because students already knew WhatsApp, the rollout didn't depend on extensive technical onboarding. Program stakeholders reported that the experience moved from planning to student use within a short window.
"Jinso brought language practice closer to the digital environment students already inhabited and created opportunities for students to practice beyond the limits of scheduled class time."
Results
Engagement results
- 94% reported engagement rate: strong participation across the registered student population.
- 40K+ messages and interactions: evidence of repeated participation, not one-time access.
- 73% deep engagement: students with more than 20 interactions during the final two weeks measured.
- #1 Host Family lesson: the learning experience with the most participation by both student count and total interactions.
Language-use indicators
Among the users included in Jinso's message-quality analysis:
- 56% increased their message length by at least 20%.
- 44% showed stronger sentence structure over time.
The analysis also identified groups of students who produced longer responses and used broader vocabulary as the deployment went on.
What Jinso surfaced
- Grammar patterns: Recurring areas for reinforcement included third-person verb forms, articles and prepositions, expressions of time, comparatives, common tense distinctions, and word order in indirect questions.
- Practical communication: Students needed more practice with polite requests, questions, daily routines, and vocabulary about family and household life.
- Sentence development: Some students started with very short or incomplete responses, which gave teachers openings to reinforce complete sentences and fuller communication.
- Engagement signals: Differences in participation helped educators see where different prompts, encouragement or extra instructional support could help.
Helping students use what they already knew
Students started the program with different levels of English. One participating student described speaking as the hardest part of learning English. Before traveling, the student used Jinso to practice pronunciation, listening and spoken communication, skills that became directly useful in conversations at school and in daily life during the exchange. The student said Jinso was especially valuable because it offered a place to practice without embarrassment. It let them repeat responses, try unfamiliar vocabulary, practice pronunciation, build longer answers, and grow more comfortable and confident before relying on English abroad.
"Regular practice helped transform speaking from the most intimidating part of English into a skill the student felt increasingly prepared to use."
Looking Ahead
From an initial deployment to scalable language readiness
Pernambuco showed that students will take up personalized English practice when it's relevant to their goals, available on familiar technology, tied to real-life situations, supported by educators, and accessible outside class hours. The next step is to make that model repeatable language-readiness infrastructure for future cohorts. Five steps toward scale:
- Represent different regional contexts. Include schools and educators from Pernambuco's varied geographic and educational settings, so the program can evaluate adoption in urban, rural, coastal and inland communities.
- Build educator preparation into deployment. Educators should understand how students access Jinso, how the learning experiences connect to program goals, how to encourage consistent practice, how to read engagement data, and when extra instructional support may be needed. Jinso could also help educators who want to build their own conversational confidence.
- Align content to program milestones. Pre-departure preparation, communicating with host families, school participation abroad, everyday problem-solving, cultural adjustment, and reintegration after returning home.
- Use data to support adoption. Aggregate data can show cohorts with low activation, groups needing implementation support, recurring language gaps, the lessons with the strongest participation, and opportunities for extra practice.
- Measure proficiency alongside engagement. Future phases could combine Jinso analytics with independently administered pre- and post-assessments to evaluate both participation and measurable changes in English proficiency over time.
Methodology & Sources
- About this case study: This case study combines aggregate platform data with anonymized qualitative evidence from stakeholders connected to Programa Ganhe o Mundo.
- Deployment period and participants: Students selected through the 2024 selection process got free access to Jinso for about four to five months, starting in September 2025.
- Quantitative data: Sourced from Jinso's internal Relatório Ganhe o Mundo 2025, based on activity recorded on the platform. "Deep engagement" was defined as completing more than 20 interactions during the final two weeks measured. Language-use indicators describe observable patterns within Jinso and are not equivalent to independently validated proficiency gains.
- Qualitative research: Jinso conducted semi-structured interviews covering three perspectives: a participating student, a public-school English teacher and a program administrator. Interview findings are contextual evidence, not a representative survey of all students or educators.
- Authorship and disclosure: This case study was prepared by Jinso Labs. Because Jinso provided the technology and wrote the study, it should be read as a documented implementation case study, not an independent academic evaluation. The deployment wasn't designed as a controlled evaluation of academic impact. All student information has been anonymized, and no student names, individual rankings, raw messages or identifiable student records are included.
