New Years Resolutions Bot

This project was a natural extension of what a MOTI-like experience could look like in digital form (why you see this project sometimes referenced as “MOTI AI”). It was my first exploration into building a chatbot, and proved a great learning experience for that type of design work. With the help of a software engineering intern who built a lightweight database and some business logic, and the GUI motion.ai for conversational flow, I put together this prototype is less than a month. to start market testing for a possible product in the space. It was a great way to start market testing for a possible product in the space, and the perfect excuse to pull in more user research on habit formation given the New Years Resolution season.

How it works

The goal with this bot was to help users set an achievable commitment for their resolutions, and then receive a custom report of how to achieve those. The report would identify if the user had enough intrinsic motivation to make it worthwhile, what element of the habit loop that user should focus on, and a guess at the style of support they should seek from friends and family for accountability.

Part of the test was a market test, to see if people would pay for this type of experience. The chatbot was packaged into a web experience with a login system for paid access. To date, the bot has had 553 unique users, including early unpaid testers.

Here is an example of the “happy path” for a user looking to exercise more:

The bot, however, could handle multiple habit types, with modules including: Exercise (with sub-modules for cardio, yoga, strengthening, walking, pushups), Reading, Writing, Hydration, Sleeping, and Other. The bot also was designed to accomodate people who didn’t know what habit they wanted to form and could make suggestions.

Key Takeaways

• Overall, the experience was very long. It’s important to make sure users know where they are in the process and what to expect. GIFs did wonders to break up chunks and give some personality and energy into the experience.

• The bot definitely caused people to reflect in ways they normally do not when setting up New Years resolutions. That was perhaps the most powerful part of the experience.

• Finding ways for the bot to fail gracefully is key. Be sure to remit any important information before moving on, so that users can verify it. There should also be a way for them to fix any mistakes.

• The cadence and timing of how quickly a bot replies does wonders for making it feel like a human interaction.