In May 2016, Facebook made an announcement that they would be investing in technology to help build chatbots for the conversational economy. Making a clear statement that the next space for innovation would not be Mobile Apps, but chatbots using SMS and Messaging platforms. At the same time many scholars and industry professionals have been raising the discussion that the API algorithms used for these chatbots would further the discrimination practices experienced in our economic sectors. Designed to favor larger more established businesses over local minority owned ones. With this in mind I had begun exploring the idea of how chatbots could help make it easier for people to track they're spending power and support their local economy.
When the Startup Institute announced that they were hosting a Hackathon in August of 2016 I decided to go and see if I could partner with like minds and explore the feasibility of my idea. I campaigned for my group to work on building a chatbot, we all really enjoyed the process and after 12 hours of work we presented our prototype and won 1st place!
In our high level planning meeting we agreed that we would use our time during the hackathon to create an MVP prototype. We discussed scope, level of effort and feasibility. Some of the more advanced functionality was moved out of scope and tagged for development in future iterations. The use of API technology, AI learning and geo-location were slated for a future version.
With our MVP we chose to work with Linear Responses vs. Conditional Logic.
We then developed a process user flow, mapping out the overall system including all of the potential actions and their coressponding message strings. We researched chatbot development tools and decided to work with Chatfuel to program the IF/Then rules that would govern our bot. Chatfuel constraints required that we pitch our MVP on the Facebook Messenger platform.
DESIGNING THE UX
Regardless of the guiding metaphor behind a bot’s personality, the basic patterns that govern its interactivity are still being worked out. We explored personality, tone of voice, sequencing of questions and on boarding. We discovered some surprising things to avoid. Such as rhetorical questions, because people expect to respond to them, even if the bot was just being polite. You would never just stick a form on your web page and not expect people to type into it.
So if bots and AIs are things that we talk to instead of "use," what exactly are we talking to? This controlling metaphor is as fundamental to a successful UX as the notion of a "desktop" is to the graphical user interface.