DIODE General

Chat as an Interface for Underserved Communities

The platform economy largely relies on smartphones to distribute work to workers (who are typically conceived of as independent ‘entrepreneurs’). In the India, many platform economy workers come from low income communities from the auto rickshaw drivers on Ola and Uber to the delivery riders of Swiggy and Zomato. In parallel to the growing platform economy, smartphone penetration has rapidly increased in low income communities in India (at least in the cities) – due to extremely cheap data and the increasing availability of economy handsets. WhatsApp is a major driver of smartphone adoption (O’Neill et al, 2016). In ICTD how to provide services – for work, health, education – to underserved communities in an appropriate way is an ongoing question. The increasing use of smartphones as a tool for work and community raises new possibilities when thinking about ICTD services. In our research, we meet plenty of auto rickshaw drivers with low or no literacy but who competently use Ola/Uber to get work and WhatsApp to maintain connectivity with their network: sharing greetings and other memes, or using voice messaging.

The almost ubiquity of chat amongst smartphone users raises exciting possibilities for thinking about chat as an interface to a whole variety of services and information designed for, with and by underserved communities.

To illustrate I will share some reflections from our recent experiences designing a FinTech service for auto-rickshaw drivers. In particular why we moved from a smartphone application to using chat as an interface. The FinTech service, Prayana, has been designed to give auto-rickshaw drivers information on how they are progressing through loans they have taken out to buy their auto-rickshaws. This service is part of an ecosystem consisting most importantly of people – field agents, drivers, back office staff and community organisers – supported by technology which helps provide timely information, accessible to those with low literacy by making use of iconography, colour coding, visualisations and numbers. The Fintech service provides information, motivators and nudges on loan progress  direct to the drivers as well as to the field agents who help manage their loans. In the field agents version it also provides workflow support for the agents.

For the drivers app we originally designed a stand-alone Android application, but our experiences with the stand-alone field agents app we had deployed in 2017 raised serious questions about the sustainability of a standalone application. The field agents app required intense development effort just to keep it running: in the 12 months after deployment of the ‘stable’ version  we released 39 versions and recorded over 900 crashes! Pretty much whenever Android updated something would break. How would our partners, Three Wheels United (TWU), a small social enterprise start-up with limited technical resources, manage to sustain it once the research was done? Let alone develop the new features necessary to keep the application in line with their rapidly growing business? Worse still, whilst we could control what phones the field agents use for their work, we have no such control over the drivers phones which consisted of a wide range of handsets and versions of android. Even building a standalone app that can work consistently across such a variety of phones is a difficult task.  Finally, a standalone app would be yet another interface for drivers to learn and in our iterative tests, even as we simplified and improved the design, this remained a barrier for many drivers.

User test of Prayana with Auto-Rickshaw Drivers in Chitradurga, India

We had been doing research in another project on how Kaizala, Microsoft’s chat app for work in emerging markets, was being adapted used by resource-constrained and innovative young enterprises to support their workflows (McGregor et al., 2019). Could this provide the solution we were looking for? The advantages of Kaizala is it works on a wide variety of Android phones, including most low end ones. Further the Kaizala team takes care of ensuring it works with new Android versions. Equally important is Kaizala’s easily programmable APIs meaning virtually anything can be built on top of it and even low resource organisations should be able to build new screens and update existing ones. And finally, the chat metaphor is easy to understand and use since the vast majority of smartphone users in India use WhatsApp. We therefore built Prayana on top of Kaizala. We used Kaizala’s backend APIs for message delivery to tailor the information delivered to each driver based on their loan performance. Visual messages are triggered to appear in the chat stream at different points in the loan, e.g. when a payment is due, has been made, or a driver is falling behind or has paid off a certain amount of their loan. Clicking on these messages opens a screen, known as a card, which gives more details about the loan in a colour-coded, visual and numeric way.

Drivers found the chat interface easy to use

The custom cards gave us a blank canvas and complete control over the way we represented information. In user tests we found that chat worked much better than we hoped and drivers had an almost instant grasp of the interface. Even though we used the same screens as we had in the standalone app, drivers understood more in this chat-based interface and were much more confident about using it. Even drivers who could not read and write understood some of the information in each card! This was very positive and an improvement on previous versions. On the downside the setup was complicated, as drivers often had never used Play Store, or did not know their passwords, and we had to set up gmail accounts for many drivers – even though they will likely never use nor need email. Some drivers did not have enough space on their phone for the app – which is likely to remain a barrier to adoption. Despite the downsides this seemed like a major improvement on a standalone app and the chat-based drivers app was deployed in January 2019.

We believe our experiences have a much wider applicability and we are currently excited about the idea of chat as an interface to underserved communities. Building on top of chat greatly reduces the barriers to building new services and apps, can be done by anyone with HTML skills and reduces the barriers for end-users adoption. We can imagine all sorts of services from healthcare to career development, from community action to workers platforms. The opportunities are immense, in India at least, where smartphone penetration is rapidly increasing even amongst low income communities. We believe chat may be a great interface to workers in the digital economy.

References

O’Neill, Jacki, Kentaro Toyama, Jay Chen, Berthel Tate, and Aysha Siddique (2016) The increasing sophistication of mobile media sharing in lower-middle-class Bangalore. In Proceedings of the Eighth International Conference on Information and Communication Technologies and Development. ACM, 2016.

McGregor, Moira, Bidwell, Nicola J., Sarangapani, Vidya, Appavoo, Jonathan and O’Neill, Jacki (2019) Talking about Chat at Work in the Global South: an Ethnographic Study of Chat Use in India and Kenya. In Proceedings of CHI, ACM, 2019

Jacki O’Neill, 21/03/2019

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s