PandaiAI

In Southeast Asia, a glaring educational gap persists, predominantly afflicting disadvantaged students who lack access to supplementary resources. While more affluent students benefit from high-quality education via after-school classes and workbooks, their less privileged counterparts face significant barriers to acquiring similar opportunities. 

Consequently, the imbalance of access to education exacerbates achievement gaps, perpetuating a cycle of social stratification and inequality. Addressing this issue demands a multifaceted approach encompassing equitable resource allocation, targeted intervention programs, and policy reforms to level the educational playing field.

To make things worse, the COVID pandemic has created a lingering problem of learning loss. Learning loss is not simply a classroom problem; it is an economic one: the Asian Development Bank estimated that learning disruption could result in lifetime earning losses of $147 billion from all Asian economies. ASEAN estimated a loss of 1.1 trillion hours of in-person learning across the Asia Pacific.

This is where AI can step in and create a large impact.

Despite AI's potential to revolutionize education, Southeast Asia grapples with a dearth of adaptive and personalized AI tools. Existing solutions in the market fall short of addressing students' diverse learning needs, failing to provide tailored educational experiences essential for mitigating educational inequality. 

Harnessing AI to its fullest capacity requires the development and implementation of innovative tools capable of adapting to individual learning styles, preferences, and abilities, as well as alignment to the national curriculum (education in many Asian countries is centralized, which follows a set curriculum).

By effectively leveraging AI technologies, Southeast Asia can bridge the education gap, fostering inclusive learning environments that empower all students to thrive academically and contribute meaningfully to society.

INTRODUCTION

PandaiAI is a comprehensive AI-assisted learning (AIAL) system that incorporates the elements of personalized, smart learning to enhance school education in the Southeast Asian region. PandaiAI aims to help develop the next generation of creative, smart and humane learners from the education system.

Using PandaiAI, students get high-quality, personalized education with AI-generated & human-moderated content, catering to all school subjects at all grade levels and aligned to the national curriculum.

PandaiAI can be thought of as an autonomous brain that can
- learn and generate educational content
- analyze students’ performance
- provide a personalized learning pathway.

COMPONENTS OF PandaiAI

These are the components of PandaiAI, with a short explanation of each:

Currently, PandaiAI has been launched in Malaysia and Brunei, integrated with the Pandai app. There are more than 800,000 users in these markets, and we plan to expand the app to the markets.

The current users are students and teachers at the primary and secondary levels. These are some of the statistics from our in-app data:

- 26% of our users come from rural areas;

- 75% of our users belong to primary level (7-12 years old);

- 92% of our active users have shown improvement after using Pandai for 2 months.

We are proud that many of our users come from rural areas, where there is a lack of access to high-quality educational interventions such as after-school tuition classes. This is possible due to two main reasons:

- the Pandai app has an offline mode where students can use it even with intermittent access to the internet;

- we work together with various NGOs, organizations and corporations in both countries to provide Pandai access (as well as mobile devices and mobile data) to students in underprivileged regions.

PandaiAI has the potential to be scaled globally to accommodate other national curricula and international standards (such as IB and IGCSE). Upon completion of development, PandaiAI is ready to be deployed, piloted, used, commercialized, and marketed to millions of school students in Southeast Asia.

PandaiAI has the potential to improve educational attainment for hundreds of millions around the world and develop thinking skills, conceptual thinking, and creativity among students (not just rote learning).

The management team has been in the education and tech sectors for more than 15 years prior to founding Pandai:

1. Suhaimi Ramly, the COO, founded an education company in 2007 and has organized more than 500 education programs in all 16 states and territories in Malaysia. These are mainly traditional education programs (in-person, paper-based, non-tech), so he had much experience on the ground, in the classroom, and dealing with various education stakeholders, such as the Ministry of Education, local educational authorities, school administrators, teachers, and students.

2. Khairul Anwar, the CEO, has been active in the international education community, especially in the field of computer science/computational thinking education. He is the national representative of Malaysia to the International Bebras Organization, a transnational educational body that promotes computational thinking education in schools. The Ministry of Education Malaysia has tapped his expertise for policy and curriculum design work.

3. Our technical team leads - Akmal Akhpah (CTO), Farahani Husni (Head of AI), and Nazatul Akhma (Head of Products) - have years of experience in tech and product development and have been involved in entrepreneurial ventures for more than 10 years before joining Pandai.

4. The Pandai team has hired a group of content matter experts in education, education technology, and AI in education, who have years of experience as educators or researchers in the field.

NB: It is worth noting that 4 of the management team members (Khairul, Suhaimi, Akmal and Nazatul) have known each other for 25 years.

- PandaiAI was launched in 2024 at the "AI for the People" launch event officiated by the Prime Minister of Malaysia.

- PandaiAI is integrated into the Pandai app, which has been operational since 2020, with over 850,000 registered users.

- Pandai graduated from Y Combinator accelerator program (Summer 2021) and has raised a US$2m seed round in 2022.

1. Access to a community of like-minded solvers from around the world, to exchange ideas and broaden our network. We have had great experience with our previous MIT Solve / Octava Foundation program. I have met all the participants in person (during a conference co-organized by Octava in Singapore). I still keep in touch with most of them, and we hope for future collaboration.

2. Access to funding opportunities: VCs, angels, funders' circle, charitable organizations, etc. to further our impact. We are not seeking to raise funds through this program specifically, but we'd like to get informed of available opportunities in the funding landscape.

3. Access to other markets.

4. Access to mentoring programs. My mentor in our previous MIT Solve/Octave program really improved my leadership practice.

PandaiAI is not just an AI tool or an AI wrapper app; it is built upon a highly validated Pandai app that has been used by 850,000 users (with 500,000,000 learning activities completed since 2020). Pandai is created with the user in mind, and we always believe that in education technology, the education part comes first before technology. 

Prior to launching PandaiAI, we populated the Pandai app with more than 1,000,000 learning content in various formats, such as quizzes, notes, videos, and experiment simulations. All the content in Pandai is aligned with the national curriculum of the markets we operate in (KSSR/KSSM in Malaysia and the BGCSE in Brunei).

The PandaiAI tool is trained using knowledge bases that are aligned with the national curricula and are available in various languages.

These are the detailed explanations of the components of PandaiAI:

PROBLEMS AND NEEDS

Problem: Lack of access to quality education for school children in SE Asia.

Cause: Technology/AI is not fully leveraged to provide quality education. No education tool offers all 3: (i) Personalized learning, (ii) Continuous learning, (iii) Meets the school curriculum.

ACTIVITIES & INPUTS

1. Designing and launching PandaiAI for personalized, continuous and curriculum-based learning
2. Signing up Pandai users through social media and school campaigns
3. Monitoring student performance.

OUTPUTS

1. PandaiAI is launched
2. Users sign up and use the app
3. Student performance is monitored
4. PandaiAI develops the necessary tools to scale.

OUTCOMES

1. PandaiAI students’ performance in the app improves
2. Rural students’ performance in the app improves
3. PandaiAI scales to other countries.

IMPACT

1. PandaiAI students’s performance in school improves
2. PandaiAI students’ performance on PISA assessments is in line with OECD averages
3. Pandai students’s performance is above national averages.

For a more detailed TOC, please refer to the following slide:

Please refer to this slide of TOC (also shared in the previous question), with the measurements/indicators shown in blue, and the impact shown at the rightmost column:

Please refer to the following slide for UN SDG target and indicator:

These are the methods, processes, and technologies used to build PandaiAI: 

1. Natural Language Processing (NLP):

AI platforms for education often use NLP to understand and analyze student queries and responses. This involves techniques such as parsing, part-of-speech tagging, named entity recognition, and sentiment analysis, among others.

2. Natural Language Generation (NLG):

AI platforms for education may use NLG to generate natural language responses to student queries or to create automated content such as explanations, summaries, or reports. This involves techniques such as template-based or rule-based NLG, as well as more advanced methods such as neural NLG.

3. Machine Learning (ML):

ML algorithms are used to build predictive models of student behavior and performance, as well as to personalize and adapt learning experiences. This involves techniques such as regression, classification, clustering, reinforcement learning, and deep learning, among others.

4. Data Analytics:

AI platforms for education often rely on data analytics to monitor and analyze student performance data, as well as to generate insights for students, teachers, and parents.

5. Learning Management Systems (LMS):

AI platforms for education often use LMS to create, manage, and deliver educational content. This involves features such as content authoring, curation, translation, and adaptation.

6. Adaptive Learning:

AI platforms for education often use adaptive learning techniques to tailor learning experiences to each individual student's needs and preferences. These techniques involve features such as personalized content recommendations, targeted feedback, and adaptive assessments, among others.

7. Gamification:

AI platforms for education often incorporate gamification techniques, such as points, badges, leaderboards, and rewards, to motivate and engage students. This involves AI principles such as intelligence game mechanics, game dynamics, and game design, among others.

8. Security & Data Protection

8(a). ISO 27001 Certification

Pandai is working to obtain ISO 27001 certification (expected to be completed in June 2024) to ensure that our solution adheres to international standards for information security management systems (ISMS). This establishes a systematic approach to managing sensitive data and addressing security risks.

8(b). Vulnerability Assessment & Penetration Testing (VAPT)

Conduct regular vulnerability assessments and penetration testing to identify vulnerabilities proactively. This helps strengthen the security posture and reduce the risk of cyber-attacks.

8(c). Encryption and Data Masking

Implement strong encryption mechanisms to protect sensitive data in transit and at rest. Additionally, use data masking to anonymize data, reducing the risk of unauthorized access.

8(d). Regular Security Audits and Monitoring

Conduct regular security audits to evaluate security posture. Implement robust monitoring solutions to detect and alert suspicious activities, unauthorized access attempts, and potential security incidents in real-time.

Full-time staff: 54

Part-time staff: 29

Contractors: 21

12 months (for PandaiAI)

Pandai has an internal ESG policy, and we take diversity of our workforce seriously. Currently, our full-time staff is more than 50% female and represents various national ethnic and religious backgrounds in Malaysia and Brunei. Pandai has also hired staff from other countries, including international students studying in Malaysia and other remote workers. We have had staff and interns from Indonesia, the Philippines, the US, Korea, Syria, Bangladesh, Italy, and other countries. 

It is worth mentioning that among the three technical team leads - two of them are females (Nazatul Akhma, Head of Product and Farahani Husni, Head of AI). Our HR policy, communicated to all incoming employees and codified in the "Pandai Staff Handbook", is committed to ensuring that everyone in the company is treated equally and with respect, with no discrimination on any basis whatsoever.

Business model: Freemium/Subscription (SAAS).

Please refer to the BMC:

1. Pandai has raised $2M from investors (Y Combinator, GFC, 500 Global, and others) in our seed round in 2021. We plan to raise our series A funding in 2025.

2. In April 2024, we received a $1M grant from the Malaysian Digital Economy Corporation (MDEC) to develop PandaiAI.

3. We have a path to profitability for the business in the next 12 months:

- steady 25% cumulative quarterly growth rate since after pandemic

- 94% average monthly retention for subscribers

- an aggressive strategy for the next 12 months focusing on 3 growth drivers: