Concept 2 - Connection with the place you are in

Created
Tags

As an Interactive journey experience

When users face the time with stress or chaos in the lifetime, they can step into the car to have a personal space while releasing their stress. However, they might have no idea about what to do and where to go to make their feeling better. As the AI recognizes a driver’s emotion (by the first concept), the AI can suggest unexpected journey plan for the driver to spend meaningful and appreciative time with the car.

📋 Our Plan - What we want to achieve, Which program we used

To suggest a personalized journey for a driver, the AI starts a conversation with the driver, saying “Seems like you need some quiet time. Do you want to go on a journey with me?” Regarding all the possibilities of the answers to the question, we drew a flow chart.

Flowchart of suggesting a journey

In order to make AI more trustworthy, we created an interaction that is more of a consensual exchange instead of biodata being constantly measured and interpreted, which feels like an invasion of privacy. The questions are vague but suggestive enough to make the user feel like they are opening up about their feelings. At the same time, they are being guided by their car to a destination that will better their current mood.

With “VoiceFlow”, a voice and chat bot tool, we created an open conversation along with ambiguous yet natural sound. The user will have the option to control the situation or have a spontaneous surprising journey.

We tested our prototype in low-fidelity context set up by using chairs on wheels and some cardboard, which can stimulate driving experience. We divided our research question in several aspects. These aspects are concept, interactions and future visions.

During the test, we needed to wizard of oz the conversation between the participant and the AI.

🧙 Exploration by prototyping tests

Findings:

  1. AI suggestions helped to make a driver in less willingness feel better from quick tests.
  1. Using a chat (typing) seems to work better than using a voice, because some pronunciations are not accurate and makes some errors
  1. The utterances should embody all possible answers, including misspellings and mispronunciations of the words.

Insights:

  1. Low yet calm voice of women (Amazon-Alexa voice) would give thrust feeling to drivers from quick tests.
  1. The more diverse utterances we input, the smarter AI responds.