Class Saathi

Class Saathi is an AI powered learning solution that empowers all stakeholders in the education ecosystem to improve learning outcomes by enabling individualized student engagement within the classroom, and adaptive learning outside the classroom. Class Saathi also generates easy to use data to drive strategy and support for teachers, administrators and parents.

It is a combination of a clicker device for each student and software for teachers, parents and administrators. Clicker devices have been proven to be effective tools in the teaching-learning process, increasing engagement significantly. After teaching a concept in a class, the teacher can ask questions using Class Saathi to gauge the understanding level of each and every student. The administrators and the parents can also see the learning data on their respective apps.

It also has the ability to work with any grade/curriculum/language and subject, depending on requirements from governments, partners, or organizations. Over 20 million questions have been solved on Class Saathi as of April, 2024.

The AI in the solution also generates rich insights into the data from assessments, and allows teachers to create question sets, as well as generate Bloom's Taxonomy aligned questions from any content given to it, in any language.

Class Saathi is aligned with the Indian National Education Policy 2020's objectives of micro-learning and continuous assessments. Our “Data First” approach ensures greater accountability at all levels. Class Saathi does not require schools to upgrade their infrastructure ("no internet" and "no electricity" are no more the limiting factors in making classrooms smart).

There are over 250 million students going to school in India, among which 130 million go to Government run public schools; and 80 million go to affordable private schools. Both these segments are often starved of basic resources, which number around 200 million students, a vast majority.

These schools typically suffer from teacher shortages and basic infrastructural deficits. According to a Central Government database called UDISE, only about 22% of schools have internet access, or any technological intervention. This means that teachers are still burdened by administrative tasks and don't have a satisfactory understanding of individual student progress, to create customized strategies. Additionally, a large number of Government run public schools and affordable private schools cater to students from low socio-economic backgrounds, where enrolment is low, dropouts are high; and teachers are extremely overwhelmed.

Class Saathi is an integrated solution that empowers teachers to teach better, administrators to understand learning outcomes of large groups of schools more, and students to learn from content that adapts to their learning levels; thus bringing transparency, accountability and visibility into the ecosystem.

The Class Saathi team is based out of South Korea and India, and consists of members spanning countries, cultures, disciplines and areas of expertise. Our team consists of both technology experts at the cutting edge and seasoned educators, teachers and social impact professionals. In particular, our team lead and the CEO comes from a small village in India and is passionate about improving lives with quality education. 

From our partners and team on ground, we are constantly collecting feedback and conducting need-analysis to build functions to empower the ecosystem even further thus building a two-way dialogue to prototype our developments before wide rollout.

Class Saathi is currently running in over 6000+ classrooms across India and South Korea, with plans to expand into more regions in Southeast Asia, Africa, and South America.

Over 500,000+ students have been impacted by Class Saathi both directly, and indirectly by non-profit organizations using it to boost learning levels in underserved communities.

Class Saathi is a unified platform that connects students, teachers and administrators to improve the learning outcomes of students, that works without internet and electricity. Primarily, it is the combination of a clicker system and software, which allows teachers to engage students, ask questions and conduct formative assessments, and view learning insights instantly within and outside the classroom. All data collected is available to view in real time through a rich dashboard, along with analyses for each stakeholder using the platform (students, parents, teachers, and administrators).

More importantly, we have built the world's first Bluetooth based clicker device, which allows it to connect to smartphones, interactive panels, computers, and run on any platform (Windows, Mac, Android, iOS, Linux). Each student is given a clicker device to respond, interact and learn in the classroom, which allows teachers to always understand their progress, and make focused learning plans that involve more than just memorization and rote learning.

Class Saathi also brings the power of adaptive and generative AI to classrooms, which allows teachers to gain insights into the learning data, as well as generate questions from any prompt, URL, or uploaded document.

It's also specifically designed to integrate and empower, rather than disrupt the current ecosystem, and is aligned to the country's education policy principles.

We have developed a new Deep Knowledge Tracing (DKT) model that has achieved state-of-the-art in the field of Knowledge Tracing, which predicts whether a student will answer the next problem correctly based on their past problem-solving data. We have written a paper (paper1) based on this technology. This model combines a Transformer Encoder with an LSTM and is versatile enough to predict not only whether the next problem will be answered correctly but also the expected solving time by simply changing the activation layer.

We are using this DKT model to precisely assess students' knowledge levels and, based on this, provide personalized homework and calculate the expected solving time for problem sets in Class Saathi.

Furthermore, we have developed a Pacemaker service that uses the DKT model to create three different levels of learning competitors tailored to the student's abilities, encouraging them to learn together and motivating them. We hold a patent (patent1) and a paper (paper3) on this, and with this paper, we won the best paper award at the 25th Electronics and Telecommunications Conference.

Additionally, we have developed an AI Tutor service that provides answers to questions tailored to the student's knowledge level. This service combines GPT and prompting technology to provide customized answers based on the student's knowledge level. We hold a patent (patent2) and a paper (paper2) related to this technology.

Thus, the DKT model and related technologies play a crucial role in providing personalized learning experiences in education, contributing to improving students' learning efficiency.

Class Saathi was tested by an independent impact assessment agency for 10 months (standard academic year) in 120 schools and 200 classrooms in government schools in 4 blocks (Cholapur, Chiraigaon, Harhua, Kashi Vidyapeeth) of the Varanasi district in Uttar Pradesh, one of the largest states in India. Over that time period, 8200+ students were totally impacted, and the following impact was seen:

• 25% increase in average attendance

• 100% student engagement

• 3X increase in learning outcomes

• 85% decrease in assessment time from teachers

Similarly, Class Saathi was implemented in 50 model government schools called CM Rise schools in the state of Madhya Pradesh, impacting 250+ classrooms and 12000+ students. From government end of year assessments, the following shifts were observed in schools using Class Saathi versus those that didn’t:

• 9.29% increase in overall scores

• 7.81% increase in Mathematics scores

• 10.25% increase in Science scores

• 9.28% increase in English scores

• 9.75% increase in Social Studies scores

Here are additional links and resources for reference:

1. Class Saathi in Varanasi - A Video

2. Class Saathi in Varanasi - Impact and Metrics

3. Class Saathi in Varanasi - Endline Report

4. Class Saathi in CM Rise Schools - Impact and Metrics

The AI in Class Saathi recognizes the learning outcomes of students by identifying strong and weak skills, and makes recommendations to teachers only based on non-identifiable information. The platform itself doesn't account for any socioeconomic data that can drive bias, and all our algorithms have been designed to be driven primarily by student responses to predefined question tags like grade, subject, topic, sub topic, skill, and more.

50 full time staff, 3 consultants

Class Saathi has already been deployed in 6000+ classrooms across India and South Korea, and is currently working with partners to expand into South Africa, Chile, Vietnam, Cambodia, and more regions around the world.

The solution is fully modular, production stock is ready to ship, and it also supports multiple languages out of the box; thus is pilot ready.

Class Saathi is implemented in public and private school networks, foundations, and organizations either through direct funding or funding through partner organizations.

We charge a one-time fee for the Class Saathi Clicker Bag, which includes support, maintenance, trainings, and content. The software included in the solution is completely free of cost. We have also built additional premium features on a subscription-based model.

For a school of 500 students, the Class Saathi Clicker Solution costs only 0.8USD per student, per year for 5 years.

Class Saathi has seen proven success in multiple countries, and is an ever evolving solution built to meet the needs of learners regardless of context. With the Learner//Meets//Future Challenge, we hope to build a better understanding of multiple learning environments to strengthen our AI model, and thus expand impact.

We would also like support in beginning conversations with additional countries to expand Class Saathi into, and get local support to tailor the solution on ground.