⇓ More from ICTworks

What About Facebook Messenger Chatbots for Development?

By Guest Writer on June 8, 2017

Bolsa Familia is the Brazilian Government program for conditional income-transfers serving 14 million families living in poverty throughout the country.

However, families have to consult the program manager at their city hall to see if they are eligible for the program. Any Brazilian poor enough to be eligible for Bolsa Familia could be discouraged from applying due to the long travel and wait times involved.

I do not mean to pick on Brazil – the problem of servicing poor and often rural constituents is not unique to that country, and delays are all too familiar to anyone who has spent any time dealing with any bureaucracy.

The question for us then becomes: Can we use instant messaging platforms and chatbots to increase convenience for everyone, especially poor people who cannot afford to waste time on tasks that can literally mean the difference between life and death?’

The Role of Chatbots

What if we could just let people find out how much they are entitled to on the Bolsa Familia by using a chat messenger? Can we create some kind of a service, powered by rules or maybe even artificial intelligence, that people can interact with via a chat interface. Is this actually possible to do?

Last year, Jacob Korenblum of Souktel wondered about the role of Facebook Messenger chatbots in development in a previous ICTworks posts.

Fast forward to today, and conversational interfaces are one of the hottest topics in technology right now. Giants like Facebook pouring resources into their development and new frameworks popping up all the time.

But is this cutting edge technology “wow” or just “meh” for us now?

3 Examples of the Wow! Factor


One of the most popular conversational interfaces of 2016 was Yeshi on Facebook Messenger .

Yeshi is a young girl in Ethiopia who walks a couple of hours every day to the nearest reliable water source. She travels alone and straps huge plastic jugs to her back so she can bring gallons of water home to her family. Chatting with her you learn about her dreams of going to school, see a map of her daily journey and can even donate money to her creators, Charity: water and Lokai.

Connecting on a human level like this opens up a range of possibilities for issue based organizations. ‘Human’ connections are powerful with audiences perhaps relating better to Yeshi than they do to the unfortunately more common images or videos of starving children. Expect more empathy interfaces showing up in your news feeds soon.


Another popular interface app on Messenger is Poncho. Local weather forecasts are undoubtedly a useful service, particularly for farmers but that is not the reason that this is interesting.

Poncho users started asking many questions, some of course about local weather conditions but also totally unrelated things like asking for a joke. The developers quickly built a feature that would tell a joke if asked.

This seems quite straightforward but is actually quite profound — in a traditional user interface, developers or designers pick which buttons to put in front of users, and hope that they in some way that they are obvious to end users.

Readers who have run any type of technology field training will be familiar with how well this can work out when its done right.  Having users just ask for what they want changes this dynamic with, potentially significant implications for our profession. We can already start to see what this could potentially mean.


Many refugees ask for translation services, to speak with doctors, aid workers, legal representatives and so on. Enter Tarjimly. This is a Facebook Messenger bot which connects volunteer translators to refugees immigrants in dire need of these services.

It was created as a result of an experience in a refugee camp for a volunteer fluent in English and Arabic who was supposed to work in operations but ended up as nearly full time translation work.  It hands off quickly to a real volunteer translator and then gets out of the way. Its success comes from giving average people a way to help those most vulnerable.

And Then the Meh Factor

I tried to actually chat with Yeshi as part of the research for this article and it seems that she is no longer available. I was informed (by chat) that the partnership formed for her creation was not a permanent collaboration and her chat has ended.  ‘Pilotitis’ is not just confined to ICT4D.

Moving swiftly on with a word of caution. Chat for any length of time and the urge will come to just reach into the screen and grab the damn thing by the neck. I promise you.

Even the famous Poncho proves less than elegant when I asked it about ice on the roads in Florianópolis. The question is not totally weather related maybe but it is not too difficult to find out the answer. For the record we have never had snow or ice in where I live.

Another pitfall is becoming locked in an infinite chat loop. This can happen when a user provides unknown context or simply does not do what the interface expects or indicates and there is no way for a user to hit Cancel or Undo Consequently the one-to-one chat experience can be rather underwhelming.

The Facebook Future of Chatbots

It is welcome news then that Facebook has just launched Chat Extensions. These allow bots in Messenger group chats and lets them function as group assistants rather than as personal friends – better matching what’s feasible with today’s technology.

For example discussions about farming in East Africa are quickly moving online, and it’s easy to imagine a bot which is market or even disease related being very useful for these groups of farmers.

Bolsa Família Bot

So, what about a chatbot assistant to help poor families in Brazil? I decided to make one.

We need to be able to reach many people so I used Facebook Messenger, as this has a wide reach in Brazil but it would actually have been better to use WhatsApp. However for now bots are not allowed on this platform.

The chat interface also needs to be quick and to provide immediate value since many people using it may be on pay-as-you-go data connections. We are not building a character but providing a service and users will not use our bot just for the sake of it — they will only use it if it provides an easier or faster way to access a service and get things done.

It should therefore give an estimation of how much the family might be awarded and what they need to be accepted on the program. The amount families actually receive depends upon a number of factors such as income level, whether there are any teenagers in the family, whether anyone is pregnant and so on.

Because of this, creating it was very interesting and very similar to creating a mobile data collection form with relevance, calculations  and grouping. Portions of chat messages can trigger another set of messages so instead of simply ensuring that all the conversation paths would work from a functional perspective, you need to test they also work from a social perspective.

There is currently no automated way to do this as far as I am aware, just in case anyone feels inspired to build such a service.

Bolsa Família Bot understands Portuguese and unlike others is a quite simple but hopefully useful service. The bot has only been live one day and already over 500 people have exchanged more than 10,000 messages with it.

Click on this link if you would like to play around with it. You’ll be taken to Facebook Messenger and asked to login. Once in Messenger, the Bot will start when you say hello.

Like any nascent industry there are many opportunities to make a real difference to peoples lives and also to earn a living doing it. If you are interested in how to do that go get our free guide on Chat Secrets

By Ian Lawrence, who helps companies create great chat experiences

DAIL Messenger Platforms for Development RFP

Editor’s Note: If you’ve read this far, you’ll be interested in DIAL’s Messenger Platforms for Development RFP that seeks a company to help answer these three questions:

  1. How are messenger platforms being used for development work?
  2. Are messenger platforms effective tools for teams deploying them in development programs, and if so which methodologies, techniques, and applications for using messengers appear most effective under what circumstances?
  3. Based on how their products are being used in development projects, what adjustments or additions should messenger providers make so that their applications are more effective for the development community?

The deadline for DIAL’s Messenger Platforms for Development RFP is Tuesday, June 13th. Please contact Carolyn Florey with any questions.

Filed Under: Featured, Software
More About: , , , , , , , , , , ,

Written by
This Guest Post is an ICTworks community knowledge-sharing effort. We actively solicit original content and search for and re-publish quality ICT-related posts we find online. Please suggest a post (even your own) to add to our collective insight.
Stay Current with ICTworksGet Regular Updates via Email

2 Comments to “What About Facebook Messenger Chatbots for Development?”

  1. Timo Luege says:

    Out of curiosity: what platform or service did you use to create your bot?

  2. Ian Lawrence says:

    I used a platform called FlowXO (https://flowxo.com/). I explain why on a follow up post on Linkedin (https://www.linkedin.com/pulse/creating-bolsa-fam%C3%ADlia-bot-ian-lawrence) which also includes stats on active user growth in the first week after launch