I conducted user research in the AI speaker domain and offered the new idea of an AI speaker calendar utility. I also designed conversational UX throughout collaboration with AI professionals.
Project Summary
Understanding
I suggested drawing an ecosystem map for the AI speaker and calendar service. I thought it was important to understand the context to expand the range of ideation for a new calendar service.
I also planned user research using various methods to gather insights that help generate ideas.
I also planned user research using various methods to gather insights that help generate ideas.
Ideation
In addition to transforming calendar applications into voice interface, this project resulted in new features that are only feasible with a voice-controlled agent.
I suggested querying for saved events with voice commands and feed-forward as the two new features that the AI speaker agent can do for calendar utility.
I suggested querying for saved events with voice commands and feed-forward as the two new features that the AI speaker agent can do for calendar utility.
Development
I designed the dialogue flow map to provide a blueprint for developing the idea into real functions.
My understandings of natural language processing (NLP) and context management technology helped me when I drew the blueprint and communicated with AI professionals and developers.
My understandings of natural language processing (NLP) and context management technology helped me when I drew the blueprint and communicated with AI professionals and developers.
Step 1
Understanding Ecosystem
In this project I designed the calendar utility of AI speaker, NAVER's WAVE. AI agents embedded in speakers help users to manage their schedules and to inquire about information of various events via voice commands.
First of all, I wanted to understand situations in which voice agents calendar utilities to users. So, I observed how users use AI speaker and drew an ecosystem of user and related services.
User Research
The next step was user research. It was to understand context and needs of users who use calendar service. In addition to observing real situations (observation and interview), I also suggested to add Role Play research. It is more generative way to research, in which researcher play a role as a AI speaker agent. I offered this method because it helps users to think beyond the current limits of AI speaker agent technology. And the insights from the research would be inspirations for designing new features of the voice agent in AI speaker.
Step 2
Insights & Concept Design
The research findings can be summarized in four key insights that led us to service ideas. The service ideas are developed into three major concepts. The following slides show the insights and the concepts.
Step 3
Designing conversational UX
The suggested concepts are separated into detail units of functions. For a AI speaker agent, each function is consist of logical flow of conversation and value-processing logic. So I designed the detail conversation scenario of AI speaker to deliver concept to real service. The scenarios are designed to provide good experience for users involved in the conversations.
The contents of this article were redesigned personally, based on materials from the project. The contents and opinions do not represent those of Naver corporation.