Problem Statement

What challenges do users face with the GE Assistant voice feature?

I observed using situations and conducted user interviews in which GE users would need a chatbot.

After research, I reframed it using Jobs to Be Done (JTBD) framework.

As a GE Appliances user. When I am cooking and have a question, I want to ask it without stopping what I'm doing, so I can get the answer and continue cooking seamlessly.

But in the current design in Smart HQ AI Assistant,

Design Explorations

Find optimal design with multi-tasking within the hypotheses

1. Accessibility of Voice Controls

Easy Button Access

Can users locate and press the voice interaction button without looking?

Accessible Interrupt Controls

Can users easily find buttons to stop voice playback or end a conversation?

2. Clarity in Voice Interaction

Clear Voice Recognition Feedback

Is it clear when the system is actively recognizing speech?

Readable Responses

Can users easily read and verify system responses on the screen?

3. Usability in Hands-Free Scenarios

Visible Command Examples

Are example commands visible and readable from a distance, such as when the phone is on a table during cooking?

1st Solution

Improved Voice Interaction with Visual Feedback

I enhanced voice interaction by enhancing visual cues, using intensified lighting during voice recognition and distinct color changes upon activation, ensuring users could easily recognize the system's status while multitasking.

1st Ideation

Interaction with Button while cooking

With a tight two-month timeline to deliver the design, I developed a user scenario focused on cooking in the kitchen. To validate the concept, I conducted experiments in the company cafeteria and proposed an initial solution based on the findings.

Usability Testing

Enhanced Interaction Efficiency and User Input Flexibility

Before the launch, I held multiple meetings with senior designers and product managers to ensure the design aligned with business goals. Due to time constraints, I conducted quick testing with employees during lunchtime, simulating cooking scenarios.

Through testing, I identified three key issues :

Interactions were limited to a small button area, making them hard to notice.

Users often provided longer responses, which were not fully accommodated by the existing input UI.

Switching between voice and chat required too many steps, making it inefficient and disruptive.

To address these issues,

I introduced auditory feedback and enlarged both the interaction area and the input UI for better visibility and usability.

Iterations

Discussed technical feasibility with engineers and Leveraging the Design System

Consult with developers to ensure the planned features and interactions are feasible, and make adjustments through ongoing communication.

Final Design

Smart HQ AI Asisstant Voice Interaction

Seamless User Experience with voice while using their home appliances.

Implemented interactions using visual clue like color, brightness to create the feeling of engaging with a AI chatbot. 

Provide Examples

Provide example questions and step-by-step guidance to familiarize users.

Noticeable Recognition of Voice Interaction

Change the lighting and text color from standby to start to indicate that it's working.

Prompting the AI Assistant

Offer one-click or voice-based prompts for effortless interactions.

Interactive Visual Cues for Distant Viewing

Implement clear indicators to confirm the AI assistant is engaged and responsive to user inputs. The interaction area has been designed with rounded edges at the top boundary of the bottom sheet for a smoother and more modern look.

Design for Longer Responses

The input field has been expanded from accommodating 2 lines to 4 lines, allowing for more detailed and engaging queries.

Optimizing Chatbot Controls for Kitchen Environments

Include a mute option for chatbot responses to accommodate the noisy kitchen settings.

Impact

Enhanced Voice Usability with High Satisfaction

After conducting testing with the final design in the company kitchen, user satisfaction turned out to be exceptionally high.

Usability Testing in the company kitchen

Launched Voice Assistant in Smart HQ with 1M+ download

After the release of the 2.0 version of Smart HQ Pro Service, customer reviews regarding device and app usability have improved. The responsive version is available across mobile, tablet, and web platforms, and it has recorded over 10K downloads.

Google Play and Appstore

Expanded to All Appliances' AI Assistant in other Smart HQ Application Series

Our voice user interaction with AI Assistant was applied in other appliances not only cooking appliances but also washer, dryer. We have expanded all the other appliances' AI Assistant.

GE Appliances Washer Detail Page

Key Takeaways

AI Anthropomorphic Interactions

I focused on designing interactions for the AI chatbot to make conversations feel as natural and empathetic as talking to a real person. The goal was to create a dialogue experience that felt responsive and relatable.

Importance of Data-Driven Design

User observation helped me design a flow optimized for cooking scenarios, ensuring practical button placement, minimal clicks, and intuitive interactions.

Designing Through Continuous Collaboration

Collaborating closely with software engineers allowed me to design detailed interactions that were both creative and technically feasible. This teamwork ensured a seamless integration of design and implementation.

Driving Engagement with User-Centered Strategies

I enjoyed studying user types and exploring strategies to keep them engaged, such as adding examples to the chatbot's home screen. It was rewarding to see these business-driven decisions reflected in improved user metrics.