Artificial intelligence is already part of daily conversations in schools. As teachers and students explore how to incorporate this technology into learning responsibly, a group of young people in the Mexican state of Guanajuato decided to use it to solve an inclusive education problem they experienced within their own school community: the lack of tools to support students with neurodivergent or mental health conditions.
“Kris Learning IA” was born from this concern, an artificial intelligence tool created by students of the College of Scientific and Technological Studies of Guanajuato (CECyTE), Plantel Coroneo. The project won second place in Solve for Tomorrow Mexico 2025.
The team members were between 16 and 18 years old and were in the final year of compulsory schooling when they developed the prototype. The idea was to use technology to support students with anxiety, attention deficit disorder (ADHD), autism, or depression who are already diagnosed by specialists.
The tool is an application that uses artificial intelligence to adapt the learning experience according to each user’s needs and provide recommendations for both students and teachers. The system is organized around three different profiles. The educational guidance area manages system information; teachers receive suggestions on how to communicate and support each student in the classroom; and students access content and activities tailored to their learning process.
Artificial intelligence proposes different pedagogical strategies, adjusts the pace of activities and uses gamification resources to promote concentration and autonomy. In addition, it incorporates a chatbot that supports students during their study and helps identify signals that may require follow-up from specialists.
In practice, the students enters the app with their “student” profile, select the subject they want to review and access study guides organized by week. From there, they can review content such as readings and videos, and in case of any doubts, rely on the artificial intelligence assistant (KrisBot) to solve exercises, summarize notes or provide explanations. As they perform activities and interact with the platform, complete quizzes, earn points and badges for their effort, and consult their calendar or progress report to keep track of their assignments.
However, the mediator teacher of the project, Ignacio Tinajero, emphasizes that the goal was never to replace psychological or psychopedagogical work. “The app is an ally. It would be irresponsible to say that it does the work of a specialist. It works together with them,” he explains.
It all started when an autistic student said that he was considering leaving because he felt he wasn’t getting the support he needed at school. That moment led the programming teacher to ask how he could help using the knowledge he had. “I came from an engineering background. I didn’t have any pedagogical tools to deal with that kind of situation,” he recalls.
His first response was to develop an open-source model that uses artificial intelligence to provide recommendations for teachers on how to interact with students with different conditions. The idea was for the app to organize the parameters of the school year, so that the teacher could visualize the progress of their classes. Through the system, they could check each student’s overall progress percentage, see their outstanding tasks, and review academic or specific educational needs diagnoses uploaded to the platform.
He brought together a small team of students with programming skills and began to build a prototype. The first student to take an interest in the initiative was named Kris. The team decided to name the project after Kris because they wanted the tool to convey a sense of closeness and identification with those who would use it.
From that reflection, they redesigned the prototype. They incorporated elements of gamification and expanded the approach to include specific strategies for other student needs. On the one hand, it works as a school activity and management app, where teachers and students monitor course progress, assignment completion, and achievements earned through points and badges. When there is a delay, the app can generate automatic “Early Alerts” on the teacher’s panel.
This way, teachers and guidance staff can intervene in a timely manner, provide close support, and connect students with professional support networks when needed.
On the other hand, it includes an interactive bot (KrisBot) that students can use every day to resolve questions about school content. It also provides a space where they can express how they are feeling during moments of anxiety or emotional distress. AI is not designed to replace the role of a psychologist or psychiatrist. Instead, at these times, it provides messages to reassure the student and guide them to seek professional help.
This new version of Kris Learning was tested with three students with different conditions: autism, Attention-Deficit/Hyperactivity Disorder (ADHD) and depression, to observe how the platform responded to different needs. These tests made it possible to incorporate specific resources, such as shorter tasks and more gamification for the student with ADHD; as well as strategies for recognition progress for the student with depression, as a way to motivate them to move forward and recognize their own achievements.
Programming was just one part of the challenge
The biggest challenge came when they discovered that they knew how to develop software, but didn’t know how to talk about neurodivergence or mental health. The teacher and students had to learn, in a few months, concepts of psychology, psychopedagogy and inclusive education that they had never encountered before.
“We didn’t know that there were different types of autism. We didn’t know the right terminology or how to refer to someone with a particular condition,’ he notes. The process involved studying diagnoses, talking with experts and understanding differences between disorders, reviewing technical language and learning how to build a tool that was respectful and empathetic.
Due to the complexity of the challenge and the diversity of needs it seeks to address, “Kris Learning IA” still needs further validation before it is ready for wider implementation. However, the tests carried out have already shown that it is possible to build a tool capable of adapting learning to different profiles and needs.
Although there are limitations, the experience has already transformed the students on the team. “A student, who still did not know what career to study, decided to turn towards the area of health to continue working with neurodivergent people”, adds the teacher, proud.
This is how they trained artificial intelligence
The development of “Kris Learning IA” did not consist solely in integrating ChatGPT into an application. The team built a specific training process for artificial intelligence to generate contextualized recommendations.
The first inputs were documents prepared by specialists describing diagnoses and care recommendations for students with anxiety, depression, ADHD or autism. The team also added observations made by teachers and the school’s educational guidance team about how each student behaves within the school environment. With these two viewpoints, the clinical and the educational, they began to feed the model.
In addition, they used research from an American university on neurodivergence as a basis for initial training of artificial intelligence. The idea, however, is that the system progressively learns from the Mexican context.
To achieve this, they established a partnership with the National System for the Integral Development of the Family (DIF) in the city where they live (Coroneo), whose specialists collaborate by providing information on different cases and variables of diagnosed students. This data collection continues as the project proceeds into its incubation phase.
The challenge of time management in STEM projects
Although the project used a structured work methodology, the daily dynamics were completely collaborative. “We didn’t want to assign a single role because we all design, we all program, we all research and we all prototype.”
The teacher was also involved in all stages of development, from research to testing and presentations. This organization allowed students to understand the project comprehensively and learn different skills during the process.
One of the biggest challenges was time management, as students were already taking eight subjects and some took up to two hours to get home after leaving school. Still, the project meetings started at night. From seven to nine in the evening, the team would connect to plan, conduct research, or share tasks. Anyone who could not connect would work on the activity assigned to them that day.
The pace was intense and the teacher decided to speak with families from the start to explain the level of commitment involved in participating in Solve for Tomorrow. “I told the parents that if at some point they saw that their children could no longer keep up, we would look for another strategy,” he said. Despite the stress, the team managed to stick together throughout the program’s process.
Another transformation was when they started preparing the STEM project pitch. Many students were afraid to speak in public. They recorded several videos, repeated presentations and constantly practiced their communication skills. “When they watch the evolution of their own videos, they surprise themselves and ask if they really spoke like that,” he says.
A project that changed the school
The recognition in Solve for Tomorrow had an immediate effect within the CECyTE Coroneo. The school began receiving invitations to innovation events and technology calls that previously did not reach the school. The students also received a prize from the program, as well as recognition from the municipal government.
“Now I’m not the one looking for equipment. It’s the students who are looking for opportunities to participate,” the teacher says. The “Kris Learning IA” team also visits high schools to share their experience and show that it is possible to develop technology projects from a small community and compete at the national level.
After national recognition, the project entered an incubation stage to become a sustainable solution. The project is being registered with the Mexican Institute of Industrial Property (IMPI) while the team develops a business model that will allow them to keep the prototype active and cover the costs of artificial intelligence licenses.
The first goal is to conduct a pilot test at 15 of the 56 plants of the CECyTE Guanajuato, working initially with five students per school. With these results, they hope to expand innovation to the rest of the state’s plants and, subsequently, bring it to CECyTE of other entities in Mexico.
“We want to continue training our artificial intelligence so that it learns how different forms of neurodivergences present themselves in different regions and can support more students,” he says.
The vision also includes adapting the tool for primary and preschool education, where many conditions begin to be identified and early intervention can have a decisive impact on the educational trajectory of girls and boys.
