Rori

Rori is a chatbot tutor harnessing AI and engaging audio to deliver personalised learning to any student on any phone

In 2020, we built a multi-award-winning solution for distance learning by radio to support children without access to the internet; We called it Rising On Air. We were able to scale it to 26 countries, through 35 partners, in 12 languages, and reached 12m children. 

We're proud to present the next step in the evolution of this innovation: Rising On Air Interactive, or 'Rori' for short.

The Problem

We are building Rori for the 617m school-age children around the world who finish primary school without having attained basic mastery of literacy and numeracy. Even before Covid-19, schools in these countries were struggling to provide students with the quality of education they deserve. Prolonged school closures have exacerbated that problem by 25% or more

The Solution

Rori is a chatbot tutor, powered by AI and sitting atop our unique library of structured curriculum content, including 500 hours of audio content covering language arts and math for five different age groups across K-12.  Delivered via SMS or WhatsApp so all students can access it, Rori will be able to pull text and audio clips from this library, personalised to the learning needs of each individual student.

Rori builds on research from interactive radio and text messaging research, as well as on the broader literature on adaptive instruction and teaching at the right level (TARL).  Rori will further contribute to research on learning-focused chatbots. 

We know that audio-based learning has been a vital tool in 2020. UNICEF found that 59% of 129 countries it studied in 2020 were using radio as part of their distance learning response. We also know that text message-based communications can be extremely impactful (Angrist et al 2020). By combining these two elements within a chatbot, in a highly personalised way, Rori will be able to advance our knowledge and understanding of these ubiquitous technologies.

Adaptive instruction, either via software or by grouping students around their readiness for different curriculum concepts rather than age, emerges from the literature as one of the most reliable ways to improve learning outcomes (Evans and Popova 2015).

There are very few chatbots in Africa and Asia that focus on learning, despite a number of successful examples in the world’s most developed markets that have been proven to improve learning outcomes (Winkler, 2018). As a result, research on chatbot technology for low-technology environments is largely missing from international education research (Rodriguez-Segura 2020). 

We will provide researchers with important insights into the efficacy of chatbots as an alternative to face-to-face tutoring, particularly in the most remote and marginalised regions where the need is greatest. 

Rori will also be able to gather learner-generated data. This data can be used to optimise the chatbot’s adaptive learning engine and help teachers to identify children in need of additional attention. Lastly, anonymized data can be shared ethically with our partners and the international research community to gain insights into the strengths and needs of learners.

Our approach is innovative in four main ways: 1) Rori will be accessible from low-cost phones, 2) students are important partners in our design process, 3) audio is embedded within the chatbot flow, and 4) we remain committed to share resources publicly.

While most learning solutions are designed for smartphones, tablets, and PCs, only ~25% of people in sub-Saharan Africa have access to these devices. Even for those using feature phones, data costs can be prohibitively expensive. Rori has the capabilities to overcome these challenges as the first AI-powered, learning-focused chatbot that works on any device.

Furthermore, we are a network of schools – not an ed-tech company. We will partner with our school communities during the pilot phase to incorporate student and parent feedback in the design. This will also keep the product's pilot overhead costs low, whilst allowing the chatbot’s recommendation engine to receive enough information to improve. 

Beyond this, Rori will benefit from the high-quality structured curriculum content we have built over the past 7 years for the African market. We are also integrating our unique audio library, developed and tested around the world in 2020. 

If Rori is successful, we will create a playbook for others to follow, in the same way that we open-sourced our Rising On Air content and shared how-to guides. We will also ensure that Rori’s data is used as a public good, informing learning engineers and researchers around the world who share our mission to provide quality education to all.

UNESCO estimates that 617m school-age children around the world finish primary school without having attained basic mastery of literacy and numeracy

These deficits, while troubling in themselves, also have compounding effects because without basic skill mastery, students cannot learn from books or other written learning materials, creating a self-perpetuating cycle where the weakest students struggle to catch up and often fall even further behind. As a result, many children leave primary school with such low literacy and math skills that further education, formal employment, and even civic participation are all but impossible.

Rori can help address all three of these challenges and accelerate the acquisition of literacy and numeracy mastery through the following activities:

1. Build and test Rori user interface and avatar experience

2. Work with parents and families to create new behaviours and daily study habits 

3. Create high-quality curriculum content to deliver via the user interface

4. Build institutional partnerships, attract users and find exciting integrations

Immediate Outputs

1. Children interact with Rori for 15-20 minutes every day as a supplement to normal school. 

2. If schools are closed, children interact between 45minutes to 1 hour per day, in addition to other educational activities.  

Longer-term Outcomes

1. Children improve mastery of reading and math (Research across the developing world shows that one of the best ways to accelerate learning for children is to teach them at the right level (Evans and Popova 2015)

2. Improved reproductive health and life chances of girls and women (Kaffenberger and Pritchett (2020) shows that the impact of time spent in education on the subsequent earnings, reproductive health and life chances of women and girls is four times higher where that education leads to a mastery of basic reading and math)

In 2021 we plan to provide first access to Rori to the 8,000 students in Rising’s 32 schools in Ghana. This will provide an immediate testbed and will ensure the product has extremely tight feedback loops with students, parents, and teachers, so as to allow a process of rapid ideation and improvement. The first phase of this pilot starts in May 2021. 

In 2022, once the product has achieved high levels of student and parent satisfaction and is showing demonstrable impacts on learning gains, we will scale to more school networks in Ghana, to Rising’s schools in Sierra Leone and Liberia, and to a number of Rising On Air partners, including school operators in Pakistan, Bangladesh, India and Nigeria. At this stage, a large scale, multi-country RCT would be delivered to rigorously assess efficacy.

In 2025, we will aim for Rori to be openly accessible to the general public and government departments of education in order to reach between 10-15m children. 

By 2030, we are aiming for Rori to have reached more than 100m children. To achieve this, we would like to make Rori free to the end-user, and accessible to all messaging platforms including WhatsApp and Facebook Messenger. We also plan to provide a freely downloadable app, on Android, iOS, and KaiOS, and work with telecommunication companies in Africa to provide dedicated toll-free phone lines.

We were awarded funding from the Schmidt Futures Tools Prize (Feb 2021) for the concept stage of Rori. Now that we have a working prototype, we are seeking funding to develop Rori through the pilot period over the next 12 months. To be successful in this stage, we will need to overcome a number of barriers: 

• The need to have an on-the-ground project lead to oversee the pilot roll-out and to ensure Rori meets relevant international data protection requirements 

• The need to ensure Rori has adverse events ‘red flag’ capabilities to establish where children might be at risk and in need of emergency support 

• The need to build a powerful avatar persona so that Rori connects emotionally with the end-user 

• The need to demonstrate to parents that Rori is a beneficial activity for their children

• The need to form agreements with the telecommunications companies to ensure costs are manageable, with little to no charge to the end-user.

Importantly, we also need to decide whether Rori can operate as an open-sourced philanthropic venture.

Looking longer-term, the priorities will be to demonstrate impact through rigorous studies and to show that Rori can translate and scale across multiple countries, in multiple languages and age groups in order for us to meet our ambitious targets. 

We are bringing together a team with a truly unique set of capabilities to build Rori. Rising brings expertise and experience in education delivery, curriculum design, pedagogy, data, evaluation, and on-the-ground know-how. Filament AI, our partner for Rori, is a market leader in the field and has built AI solutions for the likes of HSBC, L’Oreal, and the UK's NHS. Between our organisations, we have many of the capabilities to address key product development barriers.

In cases where our internal teams could use support, we will aim to lean on the collective knowledge and skills of our networks to help us work through many of these challenges. We have fantastic partners around the world in the shape of our Rising On Air community and our membership networks, such as the Global Schools Forum. We have been given extensive support from the EdTech Hub and have formed a partnership with a colleague at the University of Oxford. 

Longer-term, we will seek out research partners to design a study to assess the impact of Rori on student learning and engagement in school.

While we think there are clear pathways to commercialising Rori, we are still early in our thinking. We would like to work with Jacobs Foundation and MIT Solve to consider options for how we can keep Rori free and non-proprietary in nature so that it has the ability to grow and scale in a frictionless and inclusive way.