A slightly different, less-safe challenge.

Vessels + DFS Lab’s Chatbot Challenge

A perfect match? We hope so.

Last Wednesday, as I meandered from one blog post to another, searching for some magic bullet to help focus my efforts with Vessels, I stumbled upon DFS Lab’s Chatbot Challenge:

What is that, you say?

A challenge? (I love challenges)

Using technology to build tools for low income countries? (the very purpose of my company, Vessels)

In particular, using chatbots and virtual assistants for such purposes? (insert shameless plug for our sms chatbot service: smscollect.xyz)

Before we get too carried away…

The goal of this challenge is to find and spur on teams to create tools and services to advocate for the world’s poor. In particular, to improve the interfaces between governments and people on subsidies. These tools will take the form of modern ‘virtual assistants’ and ‘chatbots’

While chatbots might not have lived up to the Silicon Valley hype, all is not lost. Chatbots have potential in Low Income Countries as interfaces between SMS and the world wide web itself.

I have been keen to explore these applications of chatbots. In spite of the silicon valley hype, many have proven quite gimmicky (have a read of Dan Grover’s excellent summary about this). However, in places where: (1) feature phones are still popular, (2) mobile data is expensive, or (3) mobile data coverage is poor, SMS has become a reliable way for people to access tools and services.

Now add to this, the ‘power’ of artificial intelligence (more accurately, the ubiquity and low-cost of Natural Language Processing models/services), and all of a sudden, you have an interface for people to interact with new services on the internet — all using plain old conversations.

For MyWell alone, this means we can turn a complicated series of numbers and (english) characters to submit a well reading into a real conversation with a chatbot. Which is why we’re building smscollect.

Back to The Challenge

So how is this relevant to the challenge? Well, in about every way.

DFS Lab (along with a whole lot of other researchers) has found an area where chatbots and digital assistants might have potential. In their words:

Low income people are often challenged to deal with complex formal systems which are not built to their needs. These systems include banking interfaces, credit reference bureaus, KYC procedures, and especially government social safety net programs. — DFS Lab’s Digital Advocates Brief

In the brief, they place an emphasis on G2P (Government 2 Person) payments. These are payments given by governments to low income people, and usually have a list of requirements. Systems for signing up, collecting benefits and resolving issues can be difficult to understand and navigate. This is especially the case in communities with low literacy levels.

The goal of Vessels here is to:

1. learn about G2P payment services in a given country, and

2. develop a plan for integrating chatbots and digital assistants to improve these services

DFS has outlined some target countries with eligible G2P payment programs. We’re going to pick the Philippines, because:

  1. My Wife is Filipina
  2. thus, ~90% of my family is Filipino, (spot subtle Filo/Catholicism joke)
  3. It’s pretty close to home (both physically, and mentally)

This will make training the chatbot in Tagalog, understanding different resources and websites, and performing user testing and interviews rather easy.

Let the challenge begin

So let’s get to work!

I’m going to use my Medium publication to help focus my research and application. My next post will be a summary of the research into the PPPP program in the Philippines.

The team at Vessels will be working on tidying our websites and social pages a little bit, in preparation for the application.


If you like what you’ve read, make sure to click the “Recommend” button. For more pieces from myself, follow the Vessels publication, or checkout our site at vesselstech.com.