Confidential

PRODUCT: Messenger Chatbot
FIELD: Transportation
PERIOD: 2018

PROJECT

Proof of concept and workshop for a chatbot to assist new drivers of a transportation company in their registration process.
We were asked to study the possibilities of Messenger, as it was important for our partner in this project to take advantage of Facebook as an identity platform.
Facebook’s profile data together with the answers to questions we asked users along the way,  would help the brand provide a more personal experience, and would hopefully increase their driver registration process completion rate.
A chatbot was confirmed as an adequate product for this project because the organization already sends recurring communications to the users they wanted to serve with this chatbot.
Messaging channels have a higher open rate than the traditional email newsletters: 85% vs 30%, according to most studies.

PROCESS

Initially, we created the project conversational canvas, studied the users through an online research and documented user personas with the company marketing and product teams.
We did Mystery Shopping in their platform as drivers. We documented our learnings about the processes using BPMN and identified the points where the chatbot would intervene.
Then, we developed and deployed two prototypes. We used Chatfuel as a dialogue and user manager and Dialogflow as an NLP engine for recurring questions.
The two prototypes we created had the same functionality, but were expressed with different user experiences (button-based and text-based) and through two different personas named Iris and Betsy, like two characters in Taxi Driver.

irisandbetsy

We deployed both chatbots on Facebook pages created ad-hoc for this project.
We run a 4h workshop with the organization’s team. There was an all-team theoricall introduction to chatbots and conversational interfaces, the state of the art and the high level creative, development and operational processes.
Afterwards we had a smaller session with the product and development teams to go over the user personas we had previously identified, and refine them with the business insights we missed.
For the second half of the 4h session we worked with the prototypes we had created to discuss the UX, from who they users are to how to expose the brand in a channel like Facebook Messenger, that limits graphic design so much.
We spoke about style guides and about integrating brand elements in the conversation. We spoke about good ideas and bad ideas, so that the team could learn to identify them.

cabi

Next, we went through more technical details about the tools we proposed and other proprietary and open source tools available in the market to build this kind of software.
We discussed platform integrations, rules for broadcasts and other technical and product topics, includding monitoring, evaluation and training.
After the workshop we shared technical links with the team and answered open and followup questions.

OUTCOME

This project is not in production, but we are sure this will be part of the strategy of this company and many others moving forward.
For The Neon Project this was an excellent exercise to confirm the possibilities of chatbots for workflow support and reengagement.
We saw it as an opportunity to learn from the people we worked with to ultimately improve our processes and some of the tools we use in our workshops.
Do you want a chatbot for workflow support and user re-engagement for your company? Reach out to us at cristina@theneonproject.org to get started.