Beaj - Skills for Tomorrow

Educational outcomes in Pakistan are dismal: 78% of 10 year-olds can’t read and comprehend a sentence, 50% are unable to do two-digit subtraction and over 50 million students will leave K-12 education without essential skills required to succeed in the 21st century (World Bank, 2023). 2 million of these inadequately qualified graduates become teachers, half of whom have no access to any kind of professional development, reinforcing the cycle of poor learning outcomes, a poorly skilled workforce and poorly trained future teachers.

The English language is a major pain point: confused post-colonial curriculum policy requires teachers not only to teach English as a subject, but also to teach Math and Science in it. In the labour market, access to higher-paying and white-collar jobs as well as global freelance and tech markets is heavily predicated on English proficiency. However, only 6% of teachers in Pakistan are English-proficient (British Council, 2013), and they predominantly teach in elite private schools. There simply aren’t enough proficient instructors/coaches to train the remaining 94% of teachers.

Opportunities for continuous professional development are rare and incredibly difficult to access. Government programs are only available to public school teachers (and are generally low-quality) and non-profit institutions run small-scale training program for their own teachers, leaving out nearly a million under-qualified teachers at low-cost private schools across Pakistan, who struggle to teach over 20 million students.

These teachers - usually high school graduates who earn less than $3 per day - can’t access government programs, and neither they nor their low-fee schools can afford elite private programs. Local offerings are non-existent and there is a severe dearth of affordable online programs that are relevant to this context. 

Root cause: It is incredibly hard to find high quality, affordable professional development programs in Pakistan because there simply aren’t enough qualified people to train ~2 million teachers, which makes it expensive and inaccessible to all but the elite. Consequently, the majority of Pakistan’s 52 million students suffer from low learning outcomes. 

This is the enormous need that Beaj addresses. We use technology to democratize access at scale to world class instructors and relevant, high quality content through affordable courses on a mobile app, enabling teachers to bridge knowledge and skill gaps and improve the quality of teaching and learning in classrooms. 

We provide easy, affordable access to high quality, locally relevant professional development courses for teachers, starting with English language proficiency and pedagogy skills, that are developed by world-class experts and use research-backed teaching methods.  

We deliver these courses via an easy-to-use mobile app designed specifically for low-income users with low digital literacy, old android phones and patchy internet connections. A complete demo of the app is available here.

Teachers download the app from the app store, select a category and course, and can pay easily via a mobile wallet. Courses are taught in Urdu and aligned to the national curriculum and local needs. Each course is 4 weeks long, comprises 20 minutes of activities per day and costs ~$2.Regular practice and ongoing assessment is built into every course. We allocate learners to Whastapp groups to build supportive learning communities that allow interaction with fellow learners and a live instructor.

Our android app uses speech recognition technology to give real time feedback and helps users build confidence in a safe space, without the fear of judgement. We are currently integrating AI to supercharge this functionality with an LLM-linked bilingual voice chatbot that will act as a conversation buddy and personalized coach for teachers, using course material to scaffold learning, simulating real life conversations and giving feedback on how to correct errors.

Our app collects 3 types of data:

1. learner responses to practise exercises and assessments, both audio and written forms; 
2. App usage details, and 
3. User demographics 

Complementing the app, we run a Whatsapp community of +1000 teacher-changemakers to nurture emotional and mental health, build confidence and agency, and share ideas, problems and innovative solutions. We share weekly content with this group and provide a space to discuss topics and ask questions. Details and sample content available here. 

We follow inclusive human centered design principles. Our product is designed specifically for our target audience, with heavy input from teachers in our community, who test and give feedback on each piece of content and new technoloy feature before it goes live.  

Our solution has immense potential to scale. The distribution mechanism and engagement model is such that a relatively small team can deliver training programs to large numbers of teachers cost effectively, with the unit cost of each additional course decreasing as the solution scales. 

See next section for path to scale. 

Our Target Population

Our initial target audience is 1 million teachers at under-resourced ‘low-cost’ private/non-profit schools, who are under-qualified, untrained, under-confident and ill-equipped to effectively teach the +22 million K-12 students they serve. 

Over 90% of these teachers are females for whom teaching is the only “socially acceptable” job allowed by their families. Most are high school graduates with no formal teacher training who earn less than $3 a day and teach students from low- and middle-income households. 

At present, there simply aren't any affordable, high quality professional development opportunities available to these teachers. We have created this platform to provide the opportunity for continuous teacher PD through high quality courses accessible easily at affordable prices. 

Once we have established a presence in this sector, we will expand to serve teachers at public schools in Pakistan (we are already in conversations with the Federal Education Ministry which is interested in adopting our product contingent upon evidence of efficacy). 

How does our solution address specific needs?

• Our solution is specifically developed for this particular target audience. Our courses are co-designed with teachers, and every piece of content and new technology feature is extensively tested before we roll it out. 

• Our content is aligned to the national curriculum, designed for local use and taught through Urdu.  

• Our tool is easily accessible on smartphones that are widely available in Pakistan (73 million users, PTA 2023). It doesn’t require any additional hardware, making it cost-effective. 

• Our app can be used offline after an initial download, so requires minimal internet connectivity, which is a huge constraint (EdTech Hub, 2022).

• Our method is designed around teachers’ busy schedules, requiring 15-20 minutes of engagement per day at their convenience, eliminating travel costs of in-person training. This is especially beneficial for females, who have very limited mobility beyond home and school.

Theory of Change

By providing affordable access to high quality, locally relevant professional development and a community of support, we empower under-skilled teachers to bridge knowledge and skill gaps, improve mental health, gain confidence, use innovative pedagogy and unlock the potential to deliver quality education. Increased skills and agency enable them to negotiate better wages and teaching conditions. Improved teaching leads to improved learning, changing the trajectory of students’ lives with better job opportunities and higher quality of life. 

Our superpower is a unique combination of expertise, grounded in local realities, that position us for success. 

• Zainab Qureshi (Founder/CEO) has an 18-year track record in curriculum development, teacher training, large-scale program implementation, designing EdTech products for low-income users, leading rigorous RCTs, using data to inform program design and advocating for evidence-based policymaking. Years of teaching and working closely with students, teachers, school systems and policymakers has honed Zainab’s core strength: a deep, nuanced understanding of the needs, constraints, and aspirations of different actors in Pakistan’s complex education ecosystem. After 6 years of studying what doesn’t work to improve learning outcomes at Pakistan’s largest education research program housed at Harvard, she set up Beaj to try to build something that does. She uses this knowledge and extensive local networks to develop Beaj products, partner with implementing organizations and guide the team. 

• Kazi Siddiqui (Co-founder/CTO) is a technology and operations professional with experience in designing, building, and managing products and leading large engineering teams at global companies. He began his technology career building technology products for schools and low-income users in Pakistan, experience he’s using to now guide Beaj’s team of engineers. Kazi understands education settings well as he has served as a highschool Math and Physics teacher in Pakistan, which enables him to bridge gaps and design more effective products.. 

• Fizza Hasan (Content Lead) is a seasoned teacher, teacher-trainer, language acquisition expert and communication specialist. She has taught English for 29 years and French for 17 years in Pakistan. Fizza has designed our language acquisition method and is creating specialized content aligned both to international benchmarks and to the National Curriculum of Pakistan.

• Asad Liaqat (Research Partner) - (Research Scientist, Meta; HKS PhD, 2020) has been conducting applied microeconomics research for 12 years, including large-scale field experiments, belief measurement, and surveys of difficult-to-reach populations. He is designing an RCT with 4000 teachers at low-cost schools in Pakistan to evaluate the efficacy of Beaj’s English language program and its impact on various outcomes including teacher content knowledge, psychological wellbeing and wages, as well as on student learning outcomes. 

Additionally, we have a network of advisors with decades of experience doing similar work and a strong Pakistan-based team comprising 10 members.

Our Relationship with the Communities we Serve

Our community currently comprises +1000 teachers from several school systems, many of whom Zainab has known personally for over a decade. As a team of current and former teachers who have spent years studying interventions and edtech products that didn’t work because their creators didn’t understand how education and incentives play out in reality, at Beaj we ensure that every member of the team regularly visits schools, observes classrooms and engages with stakeholders. We adhere to inclusive human centered design principles and our products are co-designed with and tested by teachers from our community, whose feedback we use to iterate, improve or pivot.

We believe we are at the pilot stage, based on our outreach and usage metrics as of March 30, 2024:

• 610 unique users have begun activities (courses + free trial) on the Beaj app

• 442 monthly active users

• 347 users have enrolled in 689 courses 

• 210 users have completed 384 courses (many courses are still underway)

• We are seeing between 12 and 15 percentage point increases in users’ pre- vs. post-test scores after completing a month of instruction, indicating promising learning gains. We are working to sharpen our internal assessments and benchmark them against external assessments as we expand our user base and collect more learning data to evaluate the efficacy of our product.  

• 1004 teachers joined the Beaj ‘Teacher Leaders’ Whatsapp Community and between Dec 2023 and Feb 2024 received 12 weeks of daily pedagogy and teacher self-development content. Teachers engaged in lively discussions on topics ranging from growth mindset and social-emotional wellbeing to project based learning and effective pedagogy for English and Math. This community continues to date with weekly content and discussions, where teachers share experiences, ideas, problems and innovative solutions as they support each other. 

We are still in the process of learning from our users and tweaking our product to achieve product-market fit before we start expanding to a wider market. Each of the 10 organizations we are piloting with has said they are interested in expanding our product to more users contingent upon promising pilot results. 

Our Milestones and Timeline of Achievements:

July 2022: Beaj Education was established as an organization,

• Aug 2022 - Jan 2023: 

  • Interviewed 300 teachers, students, education leaders, parents and industry representatives to identify pain points, perceived skill gaps, and market demand

  • Ran multiple mini-experiments to identify feasible product, delivery channel and business model

  • Developed initial curriculum, built a program prototype and tested in person, digital, and hybrid delivery models

  • Designed MVP.

• Feb 2022 - Aug 2023: 

  • Built MVP of mobile app, 

  • beta tested with 90 users and iterated content and technology features.

• Sept 2023: Launched Beaj app on the Google app store.

• Oct 2023 - Feb 2024: 

  • Established Pilot Partnerships with 10 education organizations (implementation partners)

  • Upskilled 300+ teachers via courses on the app  

  • Supported 1000+ teachers through the Beaj Teacher Leaders community.

Creating social impact in hard. Creating social impact while also figuring out how to be financially sustainable is a rare feat achieved by few. Our main goal for applying to Solve is to become part of a global community of innovative, impact-driven, like-minded entrepreneurs who are wrestling with some of the hardest problems our world faces. 

We want to learn from experienced mentors who have dealt successfully with issues similar to ours, such as how to:

• use human-centered design to reduce frictions to technology adoption and create technology that can be used easily and effectively by users in under-resourced settings;
• monetize products affordably for users with low capacity to pay;
• scale products and programs without compromising quality;
• embed new/innovative products and processes into existing implementation systems that often resist change;
• achieve financial sustainability through small unit economics but large scale.

We hope to learn how to build an impactful brand, attract great talent, and make users happy. 

At Solve we hope to receive coaching and strategic advice about how to increase effectiveness, impact and scale. We also hope to learn from international best practices and cross-pollinate ideas with some of the brightest minds from across the globe.

Our solution aims to fill a vacuum in the market: lack of qualified people to train and upskill 2 million teachers in Pakistan. We aspire to leverage low-cost technology to amplify the reach of a few world class instructors producing contextually relevant content to hundreds of thousands of learners, and we want to increase the efficacy of these courses far beyond that of traditional MOOCs by personalizing them for each learner using recently available AI technology. Our LLM-linked chatbot will act as a personalized mentor for teachers, coaching them on new skills as well as on how to teach more effectively and raise student learning outcomes. 

Our solution is innovative because it applies relatively new, very sophisticated technology to benefit a population that has been historically underserved. We package this technology in an easily accessible, simple to use front end designed to bridge the digital divide by using existing, readily available hardware and accounting for low internet connectivity. There is no product resembling ours in the Pakistani market yet, and few similar products in other emerging markets around the world. 

Our solution has the potential to make English language acquisition and pedagogy training accessible to every teacher in Pakistan, which could dramatically change the education landscape, raising the current ceiling on teaching and learning quality in most classrooms. Our goal is to contribute to improved student learning at scale, particularly for language, math and noncognitive skills.

Our product also has the potential to scale horizontally beyond teachers, placing English language proficiency within reach of anyone who wants it in Pakistan, regardless of household income or educational background. This could be a complete game changer that impacts the socioeconomic landscape of the country (given that English language has long been a barrier to entry for better jobs and a tool to maintain class divides), levelling the playing field for so many and providing access to opportunities that were impossible to tap before.  

Our Theory of Change 

By providing high quality professional development we empower under-skilled teachers to bridge knowledge and skill gaps, gain exposure to pedagogical best practices and gain confidence. In turn, they are able to improve teaching quality for their students and advocate for better wages and teaching conditions for themselves. Improved teaching leads to improved student learning outcomes, changing the trajectory of students’ lives with new skills, better job opportunities and higher quality of life. 

Evidence to Support our Work

The fact that most teachers in Pakistan are not equipped with sufficient content knowledge and pedagogical skills is well-documented (Memon et al, 2010, Rashid & Mukhtar, 2012). Research demonstrates that teachers add substantial value to student test scores, particularly in English and Math, and that this value-add is significantly correlated with teachers’ own content knowledge (Bau & Das, 2020). Teachers also affect students’ noncognitive skills (Jackson, 2018). Teacher effects are found to last into students’ adult lives in the form of higher college attainment and higher lifetime earnings (Chetty et al., 2014). 

Early evidence suggests that personalised learning through AI tutors significantly enhances learners’ performance (Baillifard et al. 2023).

Numerous studies have established that higher learning outcomes impact individual lifetime earnings (Evans & Yuan, 2019). Across OECD countries, one standard deviation in literacy proficiency is associated with a 7% increase in hourly wages (OECD, 2013). In Pakistan, higher skills are correlated with higher labor market earnings, with effects being larger for women (Aslam et. al, 2010). Considerable research also shows that both cognitive and noncognitive skills affect a variety of life outcomes including employment, work experience and choice of occupation (Heckman et al. 2006).

Impact Goals:

1. Higher student learning outcomes 

2. Improved student educational experience 

3. Improved teacher content knowledge 

4. Improved teacher wellbeing (confidence, agency) 

Key Indicators and collection methods to measure progress against goals:

• Increased student learning outcomes - measured by an increase in average student test scores in select subjects, collected by an impact evaluation, most likely via a randomized controlled trial (we will begin an impact evaluation contingent upon receiving a research grant).

• Improved student educational experience - measured by a survey tool administered to students and parents to assess their satisfaction with the teacher, school and learning process (contingent upon receiving a research grant).

• Improved teacher content knowledge - measured by pre- and post- assessments for every course taken on the Beaj app. We collect all learner responses recorded on our app and closely monitor users' engagement, increase in content knowledge and improvement in other skills including speaking ability. We currently use 3 assessment tools to evaluate learner progress: an internal MCQ based in-app test, an internal oral test and an external internationally benchmarked test.

• Improved teacher wellbeing (confidence, agency) - measured by tools such as the Sense of Agency Scale (Tapal. et al. 2017) administered online.

We will also monitor impact variables like teacher wages and working conditions to assess whether these change as teachers upskill.

Specific Learning Outcomes we are Targeting & Measuring Include:

English Proficiency Skills: (taught through audio, video and text content, practiced via MCQs and speaking drills with speech-recognition technology and chatbot)

• Improved listening/reading comprehension

• Improved grammar/sentence structures

• Increased vocabulary

• Improved speaking proficiency: pronunciation, speed, fluency

Pedagogy Skills for Teachers (taught through text passages, training videos and conversation simulations with chatbot):

• Cultivating a Growth-Mindset

• Understanding the Learning Process

• Nurturing 21st Century Skills in Students

• Classroom Management, Child Psychology and Positive Discipline

• Mental Health, Social Emotional Wellbeing

The core technology that powers our solution includes:

1. Native Android app - built in Kotlin, includes audiovisual lessons, practice modules, and speech recognition technology to assess learners’ oral responses. 

2. Administrative portal (.net core/javascript/jquery) for instructors to upload course material and monitor progress via an analytics dashboard;

3. Encrypted postgre SQL Database to securely store course data and telemetry; 

4. An AI powered bilingual voice chatbot (linked to a large language model) that can a) act as a conversation buddy and personalized tutor, using course material to scaffold learning, simulating real life conversations and giving feedback on how to correct errors; and b) act as a coach and personal mentor for individual teachers, answering specific questions privately, and building confidence and agency

5. WhatsApp groups and communities to engage with learners in live ‘digital classroom’ settings We use WhatsApp because it is the most widely used application in Pakistan with 73 million active users. Telecom carriers sell extremely low-cost WhatsApp-only data packages (~$0.3 per month!) which enables such widespread use.

Our tech stack is customized for our target audience. We are attempting to take most features offline, so that users without regular internet connectivity can engage with learning material offline (after an initial download). 

Full-time staff: 7 (6 are paid, CEO is not taking any compensation yet)

Part-time staff: 4 (3 are paid, CTO is not taking any compensation yet)

Consultants on contract: 3

Interns: 3 

Active Advisors: 6 (none are paid)

1 year and 8 months: July 2022 - present

Beaj is on a mission is to create a respectful, nurturing work environment for our team, and to establish caring, supportive learning communities for our learners which empower under-served Pakistanis to build skills, confidence and agency. 

By democratizing skills that have historically only been available to the privileged, we envision a world where anyone, regardless of household income and educational background, can easily access high quality education, skill development and opportunities for upward socioeconomic mobility.

Specific activities we undertake to ensure diversity, equity and inclusion at Beaj in the Pakistani context: 

• Offer equal opportunity employment to anyone who applies, regardless of gender identity, religion, age, disability, regional or ethnic origin; 

• Ensure equal pay across genders, so as not to have a gender wage gap;

• Actively seek out applicants from universities and regions that are under-represented in the workforce;

• Allow flexibility in remote work, which is especially helpful for new parents and parents of children with disabilities;

• Even as a small startup, provide paid maternity and sick leave beyond what the market offers; 

• Actively nurture an organizational culture of inclusion, support and wellbeing, based on our core values of respect for all, extreme customer focus, continuous improvement, team support, and progress over perfection.

• Build products and offerings that are specifically designed for under-served and marginalized groups - our app is built for users with low digital literacy, really old and slow android phones, and patchy if any internet connections. Our app can be used offline after an initial download, our user interface is extremely simple and all our courses are taught through Urdu, the national language. 

• Our tech stack (Android app + Whatsapp engagement) is chosen solely based on what our users have easiest access to.

• We use inclusive Human-Centered Design to create products with and for our target communities. Our team demonstrates proximity to our target audience through regular in-person meetings and focus groups, as well as ongoing engagement through WhatsApp groups and digital communities.

• We offer high quality products at extremely low price points in order to democratize access to content knowledge and skill development that hsa historically only been available to the wealthy. 

• Our mode of delivery is especially beneficial to our female users (who make up the majority of our users) for whom mobility beyond home and school is extremely limited.

Our business model is based on distributing a digital product (courses on our app), structured as a training program that provides support to teachers while they are enrolled through live digital classrooms and learning communities that include instructors and peers. We generate revenue by charging $2 for each monthly course, which will allow us to become financially sustainable at scale. 

Key Activities:

1. Create content for courses - initial offerings include English Language proficiency and pedagogy 

2. Develop Technology - build features for android app and chatbot

3. Operations and Growth - B2B outreach, running 3-month training programs with partner organizations: onboarding, assessment, moderating digital classrooms, nudging to increase engagement and completion

4. Marketing - creating promotional material for B2B outreach and B2C outreach through social media.

Partners & Key Stakeholders

Initially, our key distribution channel is B2B partnerships with implementing organizations that employ teachers who need training. These include: 

• school systems/ independent schools, 
• colleges/ universities, 
• vocational training centers, 
• non-government organizations

We are leveraging the extensive local networks of our team, particularly our CEO, to create B2B partnerships. 

Simultaneously, we will also engage in B2C outreach by tapping individuals and networks of independent teachers through social media and teacher communities.

Key Resources Required and Cost Structure 

Money: Annual 2024 budget for current team & expenses: $100,000

Team and Cost Breakdown:

• Content development personnel + software (25%)

• Technology development personnel + hosting (45%)

• Operations and Growth personnel + travel (25%)

• Marketing personnel + social media (5%)

Once we have developed a course offering and technology features that our users love, we expect our content and technology development costs to decrease relative to the cost of growth, operations and marketing. As we scale up, the cost of delivering each unit (course) will decrease. 

Revenue 

Revenue Sources Today:

1. Initial startup capital by founders - $30,000 (29%)

2. Grant - $70,000 (68.5%)

3. BeajUrdu International revenue stream - $2,000 (2%)

4. Revenue from courses on app - $500 (0.5%)

Annual Revenue Target in 2025:

1. Grants - $100,000 (56%)

2. BeajUrdu - $20,000 (11%)

3. Courses - $60,000 (33%) based on target of 10,000 users x 3 courses each

Annual Revenue Forecast in 5 years:

1. Grants - $0 (0%)

2. BeajUrdu - $50,000 (7.7%)

3. Courses - $600,000 (92.3%) based on target of 100,000 users x 3 courses

Surplus

All profits generated will be put back into the business to scale/expand to more users. 

Path to Scale - see question below. 

--------------------------------------------------------------------------

We follow a fee-for-service business model, but tailored to a low-income target market with low capacity to play. Our unit economics are therefore very small (each montly course costs $2), but our model can be financially sustainable at scale (50,000+ users). 

We currently have 3 funding sources in addition to the startup capital our founders injected at the beginning:

1. In the short-term, we rely on incubation grants for research and development of our solution. We won an award from the Tools Competition and are finalists for an Agency Fund affiliate grant. We have raised enough funds to cover our costs for the last 1.5 years. 

2. Simultaneously, we launched a revenue-generating product (that teaches Urdu to kids of overseas Pakistanis) to cross-subsidize our work with low-income users inside Pakistan.

3. In the long run, we will rely on revenue from skill-development courses offered via our app. We currently charge ~$2 per monthly course, which is affordable but will allow financial sustainability at scale (see numbers above).

The potential to scale within Pakistan is immense. We plan to begin by targeting teachers and then expand our professional development offerings to the following underserved groups: 

• Low-cost private school teachers - ~1M, serve 20M+ students;

• Public school teachers - 1M+, serve 30M students (the Federal Education Ministry is interested in adopting our tools conditional on evidence of efficacy);

• Highschool/college/vocational center students from low-income households - 4M+

• Freelancers/technology workers - 3M+.