Eskwelabs

The business process outsourcing industry (call centers, data entry, virtual assistance) has been a pathway out of poverty and a reliable entry into the workforce for millions across the world. Today, however, many of the roles in that industry are threatened by automation, and the new roles that coming of age in the era of AI require new skillsets for which BPO workers are not prepared.

Eskwelabs empowers professionals and teachers to teach data skills to prospective learners who want to enter their industry or upskill themselves. Eskwelabs provides a rigorous curriculum, recorded video, and live lectures in which hundreds of students can participate at the same time, but channels this curriculum through on-site coaches who sit with and facilitate concept mastery with the students. This blended learning bootcamp has been pioneered with Eskwelabs in the Philippines, and is now being tested with progressively larger and more distributed audiences.

Labor-intensive low-to-medium skill jobs are the bedrock of many disadvantaged communities, but as factories and service companies alike increasingly integrate productivity-enhancing technologies of the 4th industrial revolution, hundreds of millions of these jobs are threatened to disappear - along with the gateway out of poverty that they represent.

Coding bootcamps, online courses, and other trainings exist to address this gap, but disadvantaged communities are precisely the ones least likely to succeed with these models. Cost, accessibility, ease of use, lack of community, and many other factors take us some way in explaining why the digital skills gap has only gotten bigger.

The traditional forces in sectors like BPO (data entry, customer service) are losing steam to automation, while the new roles that are booming (higher value business intelligence, engineering, and creative services) are not aligned with the skillsets present in many of these communities.

Eskwelabs works directly with call centers to help them upskill their workers into higher value-added positions, and with individuals to help them gain the new skills to enter the data industry at all levels. Eskwelabs students include call center employees, virtual assistants, data entry workers, who seek to improve their career prospects or save their jobs by advancing to more complex data analytics.

Eskwelabs also works with the employers directly, both established BPOs and rising startups, to understand the concerns and trends driving industry and to make sure that the training is aligned with what companies want now and in the future. We build data intelligence teams for BPOs, so we know exactly what BPOs are looking for in the next generation of workers.

Our solution leverages the assets that workers already bring before starting training (can use a computer, has a medium level of proficiency in English, can follow professional service norms, young and eager to learn, strong sense of community, aptitude for group-based learning and peer coaching). The solution is adapted to the challenges faced by this specific population (limited STEM proficiencies, lack of familiarity with self-paced learning, discomfort with mainly English-based content).

Eskwelabs is a solution that seeks to provide the minimum required training for someone to effectively transition into data intelligence work. This entails optimizing our product along two lines: efficacy and cost. Unfortunately, these two are often at odds - in-person coaching by a top-tier instructor with high-achieving peers is high in efficacy, but expensive and not scalable, whereas online courses are cheap but impersonal, leading to low retention, engagement, and learning. Self-paced online courses are challenging in systems where learners grew up with highly structured education systems with limited student autonomy.

So how can we manage effective learning without jeopardizing cost?

Enter blended learning. Our model combines an online curriculum and lectures that can be centrally managed (the "content") through our platform, but the delivery is managed by on-site facilitators to groups of learners (the "form"). Class consists of a mix of online video lectures (both live and pre-recorded) and online exercises with offline group work, presentations, and discussions.

From the student's perspective, our training resembles the training they are used to - there is a venue, there is a facilitator on site who explains concepts in the local language and there are peers seated around you with whom you work on exercises and group projects. All of these elements - the physical presence, the local facilitator, the peers - are the key drivers of efficacy, product appeal, student retention that we have found worked in our previous experience running a similar product for learning machine learning.

From the facilitator's perspective, our training is like a plug-and-play product that they can use to elevate themselves to teaching. The facilitators do not have to worry about curriculum, giving lectures or grading homework - their presence is needed to facilitate group discussions and project work, provide mentoring for specific topics or industry applications, and serve as motivational coach for students to keep attending class.

From Eskwelabs' perspective, our training resembles the structure of an online learning product - exercises and homework that are conducted online and whose data enables tracking of learner progress and long-term learning personalization, a top-tier instructor who provides online lectures, and similar costs to service an extra 10 or 10K students.

We are currently piloting with 100 students in 8 locations in the Philippines, and hope to improve and scale this model across other settings where communities exhibit similar challenges with the existing upskilling solutions.

There have been many attempts to bring workforce development to the public. Online courses, coding bootcamps, virtual reality trainings, Internet-based coaching, personalized learning apps are evidence of the innovation possible today.

However, these solutions have had little effect in the communities with the greatest need – communities where the educational capital and learning mindsets required to take full advantage of these solutions is lacking. The biggest predictor of the value that a data analytics course on Coursera brings to a user is probably that user's household income and prior educational background.

Our attempt at blended learning inverts the received wisdom in techno-utopian settings – that classrooms are bad and individual learning is convenient. Our founders’ experience directly running bootcamps indicates that this is often not the case. Classrooms with interactions with a real person facilitating lessons, and peers to support you, are critical to the learning of our students, many of whom have tried and failed to learn from online sources. They have grown up in an educational setting where students receive a very high level of structure and direction from the establishment, and find it difficult to adapt to a self-paced environment (doubly so with material not localized to their setting).

Our solution looks and feels like a traditional, high-quality training, combining brand-name instructors with a relatable local facilitator. The student experience takes the best of in-person tutoring, while costs & scale converge to online learning.

Evidence shows that there are significant gaps as to the reach of traditional in-person and online learning solutions to upskill workers into 21st century jobs. The experience of the employers that we have interviewed (and which have recruited from us from our data science and machine learning bootcamp) indicates that they do not find the traditional vocational education and university system in the Philippines sufficient to provide talent ready to tackle the data analytics needs of the future. This sentiment has been echoed during our preliminary market research and interviews with companies in Vietnam, Thailand, and Indonesia.

Consequently, many companies have turned to online learning or bootcamps for upskilling and recruitment. Indonesia, Southeast Asia’s largest economy, produces just 0.8 science, technology, engineering, and mathematics graduates per 1,000 citizens, trailing countries like China (3.4) or India (2).

Our own experience running data science bootcamps showed us that the blended learning model was a possibility – we experienced a 96% retention for a 10 week bootcamp over three cohorts and 85% job placement rate using the model proposed here (virtual instructor and in-person facilitators). We also have experience running a purely online course, and understand the challenges involved. Two weeks ago, we finished our 80 hour pilot for a business intelligence blended learning bootcamp with 30 students using this model, and experienced 100% retention. This is evidence to us that this may be a way to effectively provide online learning courses in Southeast Asia, and later other settings.

In Q1 of 2020, Eskwelabs is working with a corporate partner to service 100 of their call center workers across 8 different locations in upskilling into four different business intelligence roles with this product. Since its inception in May 2019, Eskwelabs has already tested this product on 75 individuals for the teaching of machine learning, and is now ready to expand this model to a more mainstream skillset.

In one year, we expect to release our product to other companies and the greater public, targeting 1,000 students reached over the next 12 months. This will help push the current understanding of the scalability of a blended learning bootcamp.

Over the next five years, we hope to become the leading provider of a new form of learning course – the blended learning bootcamp. We hope to directly reach a million workers across Southeast Asia to help them transition into a more future-proof career, and partner with many others to deploy or replicate our model in other geographies to reach millions more.

In the next year, we want to prove to the world that a blended learning model can effectively address the learning needs of one of the most crucial groups in Southeast Asia, the aspiring industrial and service worker that is expected to transition themselves and their family into the middle class. For everything that NGOs, international development agencies, or even governments have done, the reality is that it is the less glamorous factories of Vietnam and call centers of the Philippines that have taken millions out of poverty and that are driving the economic development of their entire nation.

We see our model as a demo for the world of how to create a bootcamp that can run for $50-100 per person, that has the level of standardization required to ensure consistent quality and outputs, and the level of decentralization required to have local facilitators adapt to the challenges of their particular settings and students.

In the Philippines, the largest provider of vocational education at scale is the Department of Labor’s Technical Education and Skills Development Authority. They service millions every year, but their methodology and course list is unsuited for digital skills. In every country, other actors – universities, governments, bootcamp operators, private tutors, training agencies – are wondering similar questions: what is the model that we can use to combine scale with learner effectiveness for the populations that are already on the brink of obsolescence? We hope to provide that model, for others to partner, replicate, or adapt.

Our main efforts are focused on lowering the cost of production of a bootcamp by transitioning as much as possible to an online platform, including video content and exercises, step-by-step instructions and user guides for instructors, and to reach the economies of scale required to support international instructors. However, to go further, we are also working with financing startups and recruiters to explore alternative ways for low-income students to attend, such as income share agreements or company sponsorships.

Secondly, designing such a product is expensive. Our founder team works at heavily discounted wages, but we do not expect this to be the way to attract talent to us as we scale. In particular, a true demonstration of our model would require technical and educational talent to produce something that can handle hundreds and thousands of learners.

Other barriers include the data to gradually migrate our learning exercises from fixed to adaptive, localization to different communities' educational needs, market awareness for workers and job-seekers to understand why these skills are important, legal barriers around the provision of certification or job placement partnerships, and the development of a consistent and scalable pipeline of local facilitators (due to the blended learning nature of the bootcamp, the pipeline for top-tier instructors is much easier as the number is smaller and they can be sourced internationally).

Our main efforts are focused on lowering the cost of production of a bootcamp by transitioning as much as possible to an online platform, including video content and exercises, step-by-step instructions and user guides for instructors, and to reach the economies of scale required to support international instructors. However, to go further, we are also working with financing startups and recruiters to explore alternative ways for low-income students to attend, such as income share agreements or company sponsorships. In the past, we ran our data science bootcamps as a zero-tuition up-front bootcamp through the same means, allowing students of any background to attend without worrying about not having money.

We currently generate revenue from our bootcamps which we use to grow our team. We also recruit from our top students as they combine a deep understanding of the context and setting with the student experience that we offer – we plan to continue this moving forward. We also draw on our own backgrounds and networks as founders – in San Francisco, Toronto, and Bangalore – to identify other talents to bring to the team as advisors or members. We are also seeking grant and impact funding.

Regulatory barriers are addressable in each country, although they take time and the right advisory partners. Data for AI-driven learning is a gradual process of running and scaling the model, as is market awareness and facilitator pipeline development (although we attempt to accelerate both through strategic partnerships and participation in high-visibility events and conferences).

We believe that our model works in populations for whom:

1. The evolution of digital skills is both a threat and an opportunity (e.g. current service and industrial workers at risk of automation)
2. Upskilling as provided by government or private providers is insufficient (e.g. low-income populations who can't afford it, immigrant groups for whom localization to their language is more important, non-digital natives for whom online learning is not a good option)

We plan to demonstrate the viability of our model in Southeast Asia, specifically in industries that are in flux (Business Process Outsourcing, Banking, and Healthcare) before moving on to replicate the model in other countries, in partnership with relevant actors such as ServiceNow.

3 Full-time Founders (CEO, COO, CTO)

10 FTEs (Engineering, Education, Marketing, Operations)

3 Instructors (part-time online lecturers)
13 Facilitators (part-time teaching assistants)

Our "A team" is young and fast-moving, but our backgrounds also inform a methodical and informed approach to systems change.

Angela Chen is the Founder & CEO of Eskwelabs. Angela has 8 years of experience working across investment management and sustainability at organizations like the Canada Pension Board Investment Board, WWF, and Grameen. Angela was the co-founding member of a Toronto-based impact investing advisory firm investing in emerging markets. Angela graduated from the University of Toronto, and is a Chartered Alternative Investment Analyst. Angela is a Global Shaper at the World Economic Forum since 2012.

Aurelien Chu is the COO, focused on managing processes, monitoring product efficacy, and strategic initiatives. He was formerly a management consultant with Dalberg Advisors and has worked on education reform projects with govenrments.

Anurag Adarsh is the CTO, formerly CEO & Founder, SpeakTribe & Senior Software Engineer at Google. He brings his experience as a prior entrepreneur to ensure a fail-fast approach to product development and strategy. He heads the engineering team in India.

Albert "Bash" Yumol is the lead product designer in Manila. He is an activist, data scientist, hacker, and a leader in the local tech4impact community. Bash is the founder of Mago Analytics, a data science consultancy focused on social good projects, and focuses his energy at Eskwelabs on democratizing education for the average Filipino and building the mindset for Filipinos to use their skills for nation-building projects.

Asian Institute of Management

Partnership: The top business school in the Philippines provides us with space and mentorship as the first international member of the DOST-funded Dadao Banato Incubator

Accenture

Partnership: The leading data analytics company in the Philippines sponsored our demo day and recruits from our students.

TaskUs

Partnership: A leading BPO company sponsors our data academy and recruits from our students.

Bukas

Partnership: Singaporean fintech startup pre-finances students with low-interest student tuition loans conditional upon job placement

Asian Development Bank

Partnership: The Asian Development Bank awarded us first place for their Skills for the Future Hackathon.

We partner with companies to pay for the upskilling of their staff through blended learning bootcamps, and we work with individuals to attend our bootcamp with pre-financing options by an external provider, or sponsorship by companies upon graduation.

Learners in disadvantaged communities (e.g. low-income workers displaced by automation) struggle to pay for the immediate services of an upskilling bootcamp. This is true in the Philippines, and likely holds true in other settings. This is why we seek to find those who can pay, often the companies themselves (who have incentives to fund upskilling for recruitment, staff retention, or reputational reasons, and who themselves suffer financially from the skills gap) or government (more likely to happen in ServiceNow's primary markets).

In return, learners can future-proof their career and companies gain access to a dedicated and appropriately skilled workforce. Furthermore, learners have gained a community of peers that have faced similar challenges and can support them in their continued lifelong learning journey.

We will raise socially responsible capital in order to scale in an ethical manner, while ensuring that the product is truly aligned to learner needs in disadvantaged communities.

Our product is sold to companies to upskill their current workforce (often workers that are at risk of being dismissed or automated) and we work with them to sponsor other students to enter the bootcamp (from which the companies will often seek to recruit). We are also exploring long-term partnerships with government to provide the bootcamps that they most need for the workers displaced by automation.

We currently also use cross-subsidization through selling bootcamps to middle class workers (who can pay up-front) that helps diminish the requirement to collect payment up-front for low-income students. This in turn allows us to use alternative financial arrangements that are more suited to low-income students (e.g. study-now, pay-later schemes or company placement fees).

We are applying to the Digital Workforce Challenge to secure partnerships and resources that can help us refine and replicate our model.

We want to work with ServiceNow to replicate our model in their primary markets, to expand the impact that we have had and inspire others to experiment wit our model.

We are looking for mentorships, media opportunities, and connections to potential board members and advisors that can help us expand the impact of our model. We are also open to discussions about strategic investments by impact-oriented individuals or companies. These will help us address some of the core financial and product barriers.

We are also interested in connections to US-based groups that can provide technical talent, instructors, or M&E expertise.

We believe that our model can be replicated in other settings with workers that are at-risk of automation or populations that are more threatened by the skills shift and for whom online learning has not been effective. We want to work with ServiceNow to address these populations in the markets that ServiceNow is operational in (including Southeast Asian immigrant populations).

We also believe ServiceNow can help to connect us to tech4good groups in the Bay Area and elsewhere from which we can find data scientists and engineers with which we can partner to develop the Eskwelabs model.