Travelling into the field outside the stuffy confines of a hotel conference room is always a pleasant way to break up a conference, and at the 2018 #DFS4What event, participants had the option of three excellent field visits on day two: one on psychometric testing, one on agent networks and one on understanding how chatbots are used by customers.

The chatbot field visit was organized by the Partnership for Finance in a Digital Africa with Caribou Digital, in partnership with Ker-twang and Teller, which both have significant experience building conversational interface products for emerging-market users. There is huge potential in using IVR (interactive voice response) chatbot services to bring automated voice interfaces to non-literate users. Moreover, where users are literate and using smartphones, there is potential to reach users on the platforms that they are using for transactions, such as Facebook and WhatsApp. Research by Farm.ink shows that these social media sites far outstrip formal agriculture trade platforms in terms of usage.

A telling observation from participants’ visit with women users of mobile money services was the extent to which they rely on intermediaries. Educational ability was varied, with most women having no formal education or, at best, primary education. All of them spoke of receiving text messages after receiving social payments and then having to show them to family members—usually children—before they could respond or act on the text. Given the predominance of USSD (unstructured supplementary service data) usage for mobile money services, this situation makes simple menu entries to complete a transaction tiresome and cumbersome.

What is more, all of the women kept their phone in their purse alongside a small credit-card sized plastic wallet, in which they kept their ID cards and fraying pieces of paper with phone numbers and PINs written on them. As well as using family members as intermediaries, the women spoke of handing all of these items over to mobile money agents in order to complete a transaction, which shows either a great deal of trust in agents and/or a potential security hazard.

Participants then showed the women how an IVR chatbot system works, and they all immediately responded to a voice call telling them they had received some money after a transfer was made to their account. Of course, part of their positive response was because of the good news! However, it was also because they could instantly understand what was happening. Some women initially said they would be unlikely to answer a call from an unidentified number, but then they said they trusted the IVR system once they knew it was legitimate. Even better, they realized that the IVR system would enable them to recall their PIN number via the options, meaning they no longer needed the insecure and decaying pieces of paper they carried around with them.

The visit provided interesting insight into the way chatbots can change interfaces for the better for low-income and low-literacy users. Trust remains an issue, and there is clearly a need for face-to-face interaction to run through a new IVR chatbot system at roll-out. Yet, it seems that once trust is established in voice as an interface, powerful new capabilities of users are unlocked.

December 2018. Copyright © UN Capital Development Fund. All rights reserved.

The views expressed in this publication are those of the author(s) and do not necessarily represent the views of UNCDF, the United Nations or any of its affiliated organizations or its Member States.

Mention of any firm or licensed process does not imply endorsement by the United Nations