The way we interact with technology is undergoing a profound transformation. The rise of natural language interfaces (NLIs) marks a significant shift from traditional graphical user interfaces (GUIs) toward more conversational and intuitive systems. Drawing from my experiences at rabbit inc., where we are shaping the future of natural human-machine interaction, this article explores the evolution of interface systems, some UX practices from my work on rabbit OS, and what we can expect moving forward.

banner credit: luma.ai

Observations: A Shift from guis to vuis

Today, people increasingly turn to specialized platforms to manage their information and tasks. For example, platforms like pinterest serves as a source of inspiration, while chatbots are handy to help with day-to-day tasks. Traditional search engines, while still relevant, are losing their dominance in providing results that match with peoples’ intention. As people increasingly use their phones for entertainment—video calls, mobile gaming, and watching videos—the opportunity arises to redefine how we handle information and tasks.

What’s been prevailing in the past decade is Graphic User Interface (GUIs). The dominance of GUIs can be traced back to the early days of personal computing. GUIs made technology more accessible by allowing users to interact through visual elements like icons and menus. However, as technology evolved, so did user expectations. The introduction of natural language processing (NLP) in the 1960s laid the groundwork for voice-based interactions, but it wasn’t until the AI surge of the 2000s and 2010s—with innovations like Siri and Alexa—that significant advancements in text-to-speech (TTS) and automatic speech recognition (ASR) were achieved. These technologies have transformed from rudimentary text-based commands to sophisticated voice and text interactions. The resulting decrease in latency and improvement in accuracy make voice user interface a compelling alternative to graphic user interface.

early assumptions

As we look to the future of NLP interfaces, several assumptions can guide our expectations:

1. Beyond GUIs: While GUIs have their strengths, they are not always the most efficient way to achieve user goals. With AI and NLP, we can streamline tasks, especially when users have a clear intent and seek quick assistance. For instance, in scenarios where users need immediate help or information, NLP can provide a faster and more intuitive solution than traditional interfaces.

2. Technological Readiness: NLP technology is now mature enough to be applied in real-world scenarios. The continuous advancements in AI mean that NLP systems can handle complex interactions and provide accurate, contextually relevant responses. This readiness paves the way for more widespread adoption of voice-first and conversational interfaces.

3. Evolving User Behaviors: Users’ behaviors and preferences are constantly evolving, particularly with mobile devices and service platforms. This shift lowers the learning curve for adopting new ways of interacting with technology, making it easier for users to embrace NLP-based interfaces.

Designing for NLP-Oriented Interfaces

When designing NLP-oriented interfaces, several methods and insights are crucial:

1. Understanding User Needs: To design effective NLP interfaces, it’s essential to understand how users will benefit from the system. Identify the scenarios in which users will interact with the device and predict their intentions. For instance, in Rabbit OS, we focus on scenarios where users need quick, contextual assistance and tailor the interface to meet those needs.

2. Design with Empathy: Empathy is at the core of designing accessible and user-friendly interfaces. Ensure that your NLP interface accommodates diverse user needs, including those with disabilities. This involves creating a system that is intuitive and provides multiple modes of interaction, such as voice and text.

3. Scalable Conversation Components: As NLP systems handle increasingly complex interactions, designing conversation-related components that can scale is crucial. This means creating a flexible and adaptive system that can handle various types of conversations and continuously improve based on user interactions.