Project Melo is a voice assistant that learns and adapts to what users need through conversation and provides tailored information and recommendation. Melo is aimed to increase sentimental attachment and increase motivation by natural conversation and values it brings to users.

In the project, I led the research process and translated findings into insights. I was in charge of creating wireframes and initial onboarding experience. I created low-fi and high-fi prototypes.

Nov 2018 - Dec 2018
AI, Voice Assistant

The landscape

Discovering the Problem

During our research, we conducted 5 interviews with the current voice assistant users to understand the landscape and how people talk to each other when they meet for the first time.

Weak Motivations

Mainly use the assistant for hands-free interaction, but there's more that can be done without using the technology.

Not Personalized

If users ask for some information, the assistant directs user to either different apps or brings up a general information.

Unnatural Interaction

While interacting with the voice assistant, it is hard to know when the conversation ends.

Identifying Design Goals

Natural Conversation
The conversation should feel natural so that users don't feel awkward to have conversation.
Non-Verbal Cues
People naturally communicate using non-verbal cues like facial expression or body postures.
Voice assistant should adapt to user’s way of interacting and change its behavior.

leveling up
the Interaction

Introducing Melo
Please turn on the sound

First Impression Matters

Introducing the personal voice assistant should feel organic as it is with getting to know new people. The name of the assistant and the look and feel are initialized through user's voice.

Adapts to User's Preference

Every Melo looks, talks, and behaves differently, reflecting on how the user interacts with the assistant.

When the user first interacts, Melo will look more generic, and won't have a specific personality.

But as Melo learns more about the user, the appearance and the characteristics will change over time.

Calm, Empathetic, Modest Melo

Curious, Optimistic, Melo

Enthusiastic, Funny Melo

Tailored Information

In the current Siri or Google Assistant, when users ask for a weather, they will receive a series of numbers like what the actual temperature, humidity...etc are.

But what if Melo can provide more? For instance, Melo can analyze user's photo album to display what they wore on a similar weather and provide a more valuable and contextualized information.

Understand's User's Mood and Conditions

Melo can understand user's sentiments, mood, and condition through detecting cues like the tone of voice and the speed of user's response. Melo will try to relieve users' concerns like their best friend.


Understanding How We Talk

During our research, we conducted 5 interviews with the current voice assistant users to understand the landscape and how people talk to each other when they meet for the first time.

Nature of conversation

We discovered that there has to be a seamless transition in taking turns, a clear indicator of the ending and discovery of a common topic to facilitate a comfortable conversation

Conversing with an assistant

From testing, we learned that setting clearer expectations to the user about the onboarding process is crucial. Altering the assistant's voice didn't feel like a natural introduction to the assistant's personalization capabilities

Initial Ideation

The first design starts an avatar coming out of a box when called upon.
But based on testing, there needed to be more fun and interactive encounter

Iterating a Concept

The second design morphs into a character every time the name is called.
This way, users can feel the sense of creation compare to something that just appear.

Question about the design? Let's Chat