Astrid

Voice based English language private tutoring system powered by AI, providing affordable and easy access to top-tier language learning.

Two billion people are actively learning English as a life-transforming skill, arguably one of the largest learning cohorts in the world. 

Unaffordability and lack of teacher access prevent many from progressing beyond acquisition of basic English phrases and concepts, excluding them from educational and professional opportunities (formal and informal), and hampering life-chances and livelihood improvement. The result, for many, can be remaining in/falling into poverty and the inability to realise their full potential, especially so for girls and women. 

This is unacceptable - and Astrid has the solution. For us, communication skills, voice and the ability to express ideas is central to practical language-learning. Astrid makes this possible for learners of any age, anywhere at any time, empowering parents, teachers and students, indeed any learner.

Built mobile-first and available for Android and iOS, Astrid listens, evaluates, actively supports and coaches English pronunciation and fluency practice in real-time. It collects actionable insights based on data and performance that generate reading, comprehension, pronunciation and speaking exercises tailored to each student. 

All the student needs is a voice. 

Education is key to a person’s life chances - employment, mobility and the future prospects of the individual and family. Along healthcare, education is a fundamental driver of well-being and vital change-driver. It also builds social capital, undergirds economic development, improves health outcomes and is critical to wealth accumulation and emergence from poverty. 

English is rightly viewed as a crucial skill by most governments, educators, parents and employers worldwide. South-East Asia is no exception. As the region evolves away from low-value, high-volume output, towards higher value, more service-orientated offerings, clear basic to fluent English is more critical than ever for more middle- and low-wage earners. Indeed, the opportunity cost of not commanding English will become a deeper socio-economic divider and excluder every day, especially for those whose parents cannot afford expensive, often private local solutions. This leaves most learners with the public school system of variable to poor quality, or no English instruction at all.

Further, teacher shortages disproportionally affect developing countries and the poor, especially if alternatives to in-class instruction are unavailable or unaffordable. The global pandemic has put this in stark relief, decimating many of the affordable private offline solutions that have emerged in recent years.

To demonstrate the supply-demand asymmetry, the recent announcement by Thailand’s Ministry of Education to recruit 10,000 native English-speaking teachers can be juxtaposed with UNESCO's estimated deficit of 70 million new teachers needed globally to meet SDG 4 by 2030. It's obvious that technology is a sine qua non for closing the gap.

The Astrid solution supports both young learners (initial focus ages 5-18) and parents, whilst achieving a ‘multiplier effect’ for teachers (already in limited supply), who easily can incorporate Astrid into everyday teaching. 

Though teaching quality in South-East Asia is relatively strong compared to many analogous developing regions, access and cost remain potent excluders, (especially in the two archipelagic nations, Indonesia and Philippines). Further, the grammar-based, often exam-focused, repetition/rote-dependent pedagogic approach followed in many schools impedes practice, confidence-gaining, conversational prowess, listening and comprehension skills. 

Particular challenges in South-East Asia related to pronunciation and access which further disadvantages learners compared to peers elsewhere. Currently the best, if still limited, solutions are expensive, off-line, generally urban-centric and hence available only to the top tiers of the economic pyramid. 

Astrid turns this paradigm on its head by making personalised English-speaking and conversation practice available for all. It does so by delivering non-grammar-focused, voice-based learning, tailored to each learner.

The previous careers of Astrid’s key staff range from building some of the world's largest learning platforms to development finance, private-sector development and inclusive business focused on base-of-the-pyramid populations in Asia. This combination gives Astrid unique insights into engaging lower-income, excluded social segments. It has also taught us that the oft-repeated phrase in development/impact circles, "democratising access" can be a fig leaf for dumbed-downed product solutions that fail to take account of the high price/quality-elasticity of lower-income groups, especially related to education and healthcare expenditure. In other words, Astrid recognises that education and specifically English acquisition is a significant proportion of many South-East Asian households' monthly expenditure. Hence, tangible, rapid improvements at any age are critical, both to Astrid’s business model, and to realising its impact for beneficiaries. 

We founded Astrid as a commercial response to a pervasive problem and socio-economic divider: English competency and fluency. Inclusion and impact are intrinsic to our model because in order to scale, Astrid must perforce optimise its price for the less well-off; this is to say, it is our goal of ubiquity via concrete English language improvement that will drive commercial success, because mainstreaming Astrid necessarily entails incorporating lower segments of the economic pyramid, and it is in such segments that the most pernicious ‘externalities’ of absent English skills are found. 

By leveraging technology, we can level the playing field for young girls and boys, and women and men, enabling them to improve their lives and to compete in Vietnam, Indonesia, Philippines, Thailand, Malaysia and beyond.

Founder and CEO - Andreas Kullberg

We have developed an innovative, unique model underpinned by proprietary technology to deliver the gold standard of English language learning (1-on-1 tutoring) to all. The model is directly enabled by our unique speech recognition and data-driven learning platform. Furthermore, it's purpose-built to fortify the role of teachers, not replace them.

The technology provides an AI-powered language tutor for self study but as importantly we leverage the platform's data analysis and automation to enable accredited teachers to provide private tutoring at a drastically lower price point. This enables the creation of a world class learning journey for each learner, combining the best of AI and human instruction. 

Moreover, our tech team continues to push boundaries by applying recent breakthroughs in machine learning, to current and forthcoming aspects of Astrid to ensure continuous innovation. One example, especially relevant to this Challenge, is our development of self-writing books/exercises authored by AI that are indistinguishable from human-authored texts. This creates endlessly scalable libraries of relevant, suitable and high quality content at zero content cost. AI-created texts (using GPT-3 and other libraries) combined with our proprietary automated CEFR-levelling tool ensures academic alignment and scoring. 

Our self-writing books embodies our idea of the Astrid "magical schoolbook", which adapts to cover all needs for all students, a versatile, engaging tool adapted to each student throughout every learning journey.

We launched our solution/app in March of this year, and currently have learners in 140 countries.

In addition we have worked with schools and teachers around the world for nearly 2 years to develop our product, providing Astrid to them to ensure continuity of English language learning during school closures necessitated by the pandemic.

To ensure third-party validation from a rigorous academic/pedagogical perspective, and after very strong learning outcomes measured in partnership schools using Astrid, we launched a project with Yale University to study learner progress and instruction efficacy in October 2021.

Astrid is voice-based and personalised at its core, built around our proprietary speech recognition (Astrid Voice) specifically developed for language learning. Astrid Voice uses state-of-the-art machine learning to provide pronunciation feedback in real time.

We score pronunciation, fluency, etc. and provide personalised feedback and guidance, while adapting the learning journey based on learning gaps.

Astrid Voice is at the very fore-front of accuracy, both on a phoneme and word-level basis, while keeping low latency times. The precision of the Astrid Voice's real-time feedback has a strong motivational effect on learners to further boost engagement and progress.

Astrid has developed a proprietary AI powered speech recognition system that mimics a human private tutor – what so many parents desire for children but is inaccessible/unaffordable – and produces the following unique output: a tailored learning-journey for each learner that focuses practice on specific aspects of pronunciation, comprehension and listening skills, doubling down on areas of weakness. This generates learning data that drives Astrid’s second key component: a unique private-tutoring function enabling teachers to hone in and provide feedback to scores of students. 

The contrast with in-class teaching (where/when available, of decent quality– all problematic in the region) is that class sizes and single-teacher/teacher-capability levels impede practice opportunity, maintain or boost student inhibitions, and discourage individual confidence practice. The outcomes achieved by disrupting this vicious circle include far faster, more accurate, practicable acquisition of speaking, listening and comprehension skills that accelerate the learning journey. 

The individual, household socio-economic and societal impacts, highlighted above, range from reduced exclusion and inequality based on income, geography and gender (among others), to greater livelihood security due to greater income-generating opportunities, whether formal, semi-formal or informal. 

Hence, Astrid’s theory of change is that affordable, engaging and, critically, pedagogically-effective tailored language tuition powered by AI vastly accelerates English language competency-to-fluency for all social segments in all locations.  

Astrid is a data-driven company, whose solution is itself predicated on improved English-learning outcomes. Whilst in time, we will be able to track the ‘livelihood delta’ conferred by English competency/fluency, the focused indicators currently tracked include reading, speaking, pronunciation score, fluency and overall CEFR English ability levels. In addition to proficiency indicators, Astrid also tracks and provides insights on learner Efforts in order to provide a balanced and nuanced learner profile.

The relationship between these indicators and overall progress on impact goals is symbiotic. If English acquisition is fundamental to educational, professional and other life prospects, then as our data sets grow and cohorts progress, we will be able to track associated key indicators such as educational outcomes, securing of tertiary and vocational education places, gaining of employment and the ‘job quality delta’ related to English-speaking skills, mobility in work, contribution to entrepreneurship and self-employment, ability to engage/work within the digital/virtual/internet world and so on. 

We have the following core impact goals in 2021-2026. 

First, to reach 1m+ users in Vietnam, beyond the main urban centres of Hanoi, Ho Chi Minh City and Da Nang, and into rural areas. 

Second, to do so with unwavering focus on affordability and access, achieved through economies of scale and innovative financing solutions for lower-income groups that accommodate cash flow volatilities and income limitations – (we are currently in discussions with microfinance partners and telecoms operators to this end). This will facilitate penetration of lower-tier cities and rural areas. 

Third, to leverage the success in Vietnam to expand to Indonesia, Philippines and beyond. In addition to this largely 5-years onwards age focus, we believe, fourthly, that it will be critical to penetrate the tertiary, vocational, young-professional and continuing professional/adult education segments. This is because SDG 4 achievement and national development goals will only be possible through skills and workforce upgrading which, in turn, is often a function of English competency. 

Though highly ambitious, these goals are eminently achievable because the Astrid solution is inherently scalable and replicable by market, learning environment and cohort profile (youth, young-adult, professional etc.). We also believe our unwavering focus on practical English skills-acquisition through individual practice and tailored drilling that is engaging and fun is key to customer acquisition, retention and rapid further growth. There is a reason why, in just 6 months' time Astrid is already being used in 140 countries worldwide. 

We are exploring distribution partnerships for our proven solution across all of the countries covered by this Challenge.

Strategic partnerships are a key pillar in Astrid's distribution strategy. We are currently in advanced discussions with mobile operators, banks and other consumer-facing businesses to penetrate all social segments in the region and to reach scale.

Astrid is a natural partner to many customer-facing businesses, whether via mobile subscriptions or savings accounts, because it addresses a core household priority requiring mobile telephony and disposable income. In time, Astrid will also be positioned as a vital training tool for businesses and employers, to accelerate staff proficiency in English and move up the value chain.

The founding team at Astrid previously held senior leadership positions at some of the world's largest learning platforms (EF), consumer/media (Spotify) and gaming (Candy Crush) franchises. We built some of the most popular digital apps and solutions in the world and realised we had insights and experiences that we could leverage to deliver positive social impact at scale. Added to this was our experience as highly active impact investors in emerging markets which strengthened our conviction that the stars were aligned for education to adopt true digitalisation. This was in 2019.

We also shared a strong view that "democratising access to education" should _not_ mean offering inferior products at a lower price point. Instead we felt it was paramount if you wanted to drive real sustainable change that you had to offer the absolute best level quality to everyone in new innovative ways. That's how you level the playing field.

The team currently consists of 20 people across Product, Tech/Machine Learning, Academics and Design.

https://www.astrideducation.com/about-us

The Astrid team consists of experienced engineers, academics, and company builders with backgrounds ranging from EF, Google, Spotify and King. 

We have a proven track record of scaling digital education products to schools and consumers globally, world class technology and AI domain expertise combined with extensive experience and networks in our markets.

Our uniqueness is not merely a factor of our strong technology and product, but more so the diverse backgrounds and experiences which we combined to build our solution. We have deep on the ground experience of building and driving positive social impact across some of the most challenging markets in Asia and Africa. From driving financial inclusion in Cambodia to addressing access to quality healthcare in post-conflict areas in Sierra Leone, from designing, launching and managing a USD20m women empowerment initiative across global developing markets to building EdTech products with teachers inside classrooms that can offer premium level education to the base of the pyramid in India. From winning all categories in Duolingo’s global AI competition to creating some of the most popular mobile games enjoyed by hundreds of millions every month. From revolutionising the healthcare system in Europe to being parents and profoundly understanding the universal hope of seeing ones children grow up with the best possible future. 

With all these experiences, we come together as a team that firmly believes we have a real shot of helping millions and millions of people around the world. To us, it would be unthinkable not to pursue this mission.

With over a decade of experience in developing digital learning solutions, our Team Lead pulls from his own experience with similar tasks, evaluating what might work best, making the decision and being confident in taking the responsibility for the result. 

This has been key to the successful development of our proprietary Private Tutoring model, where he led the product team through a series of research, evaluation and problem-solving sessions that resulted in a strong and collaborative team of individuals working toward the same goal. 

Our unique Private Tutoring model addresses the issue that learning outcomes decrease dramatically when EdTech providers sacrifice teaching quality and personalized remediation in order to deliver tutoring at low prices with high margins. From his previous experience of building tutoring platforms internationally for both ELT and K12, our Team Lead could decisively articulate his product vision to the team and enable them to develop a private tutoring solution that combines high academic rigour and certified native English teachers with an affordable price point that reach the larger population.

Some of the partnerships we have in place, include: 

Oxford University Press - focused on research, combining their 150 years of excellence in language research with Astrid's technology to continue develop our solutions for learners. https://medium.com/astrid-education/astrid-education-partners-with-oxford-languages-part-of-oxford-university-press-6d19a9bacad2

Yale University - currently launching an efficacy study together with researchers from Yale's linguistics department

Waterbear Network - a sustainability-focused content partnership with Waterbear; the first interactive streaming platform dedicated to the future of our planet. The collaboration will see WaterBear feature its high-quality content to young learners around the world on Astrid’s learning platform. https://medium.com/astrid-education/astrid-education-signs-sustainability-focused-content-deal-with-waterbear-network-90ade1c61ba2

Centre for Innovation Japan - The CFI is a not-for-profit organization devoted to supporting the Japanese Ministry of Education’s (MEXT) recent reform initiative, aiming to develop student-centered, inquiry-based, collaborative, and internationally-minded approaches for next-generation Japanese learners. The partnership with Astrid is focused on support new forms of learning, designed to empower both teachers and young learners in one of the most digitally advanced countries in the world. https://medium.com/astrid-education/astrid-education-partners-with-cfi-in-japan-to-contribute-to-japans-educational-vision-using-bf2050be8372

BuffaloGrid - BuffaloGrid offers a solar powered hub to deliver digital services for the unconnected around the world. Together we explores a solution to increase our reach to rural areas where electricity and especially internet connectivity is a real issue.

When we were informed about the Innovation Challenge directly from the Octava Foundation, it was genuinely the first time we saw an initiative that appeared to be fully and completely aligned with our focus in terms of solution, impact aims and geography.

We believe there's an exciting partnership opportunity that can be built on combining Astrid's solution and Octava's extensive local/regional market insights and network, designed to scale impact in the region.

In addition, as we build, develop and refine our impact framework and methodology we view a potential partnership as a fantastic opportunity to benchmark and guide our work.

As we grow in the region, we want and need to take great care in how we navigate the local legal and regulatory environment. This can be prohibitively difficult for a start up.

As a solution provider who's very comfortable in its product/solution, technology and academics, while having seen fantastic user feedback and engagement, we're mindful that our greatest challenge is distribution to scale the user base.