2025 UX Trends — The Evolution from UI (User Interface) to UI (User Intent): A New Era in Human-Agent Interaction
As we enter 2025, user experience design is undergoing a significant evolution, bringing breakthrough changes in human-machine interaction patterns. Over the past decades of traditional human-computer interaction development, people primarily interacted with computer systems through Command Line Interface (CLI) and Graphical User Interface (GUI). Users had to learn and follow specific operational procedures established by designers, such as memorizing command syntax, understanding icon meanings, and following operational sequences, while computer systems executed tasks according to these explicit instructions. Designers’ work during this phase focused on ensuring users could understand how to operate computer systems, including interface usability, learnability, and visual experience. However, after decades of development, these so-called UI/UX design processes have matured into a fairly standardized design workflow, with companies establishing design specifications for interaction modes, resulting in nearly identical operation methods for existing Web and App interfaces, allowing people to intuitively operate and understand system functions.
Subsequently, with the advancement of artificial intelligence technology, we have now entered an era of more diverse human-machine interaction modes. The most distinctive feature of this stage is the emergence of multi-modal interfaces, which not only retain traditional graphical interfaces but also incorporate natural interaction methods such as voice and gestures, allowing users to interact with systems more intuitively. Through AI technology assistance, systems have begun to develop preliminary abilities to understand user intent, such as voice assistants comprehending natural language commands and gesture recognition systems understanding user movement intentions. Systems at this stage are a combination of traditional computer systems and AI technology, capable of not only executing explicit commands but also understanding and responding to more complex user needs. The advantage of multi-modal interfaces lies in providing multiple interaction choices, allowing users to select the most suitable interaction method based on context.
Moving forward, we are about to welcome a series of emerging human-machine interaction stages, as user experience is undergoing a fundamental transformation: shifting from focusing on “how to operate” to understanding “what is wanted.” In this stage, the interface layer becomes more transparent, allowing users to express their needs directly in human language, while AI Agents equipped with multi-modal and highly empathetic understanding capabilities can deeply comprehend user intent and proactively take over, cross-platform, and even time-share and synchronously plan and execute solutions. This “direct communication” interaction mode makes human-machine interaction more like a thought exchange, where users can iteratively express their true needs and goals faithfully, while AI Agents are responsible for understanding intent and planning execution details. For example, users can directly express “I want to build an energy-efficient smart home system for my new house,” and AI Agents will understand this high-level goal, consider the user’s living habits, budget, and environmental factors, proactively propose complete solutions, and continuously optimize and adjust during execution.
Notably, these three stages do not replace each other but continue to stack and complement one another. As shown in Figure 1, each new stage adds new capabilities on the foundation of the previous one. In different usage scenarios, users may need different interaction modes: sometimes requiring GUI’s precise control, sometimes preferring the convenience of voice commands, and sometimes needing AI Agents’ deep understanding and human empowerment for automatic execution. This multi-level interaction architecture creates a fluid and natural user experience, transcending the original limitations of human tool operation interaction, enabling technology to actively understand humans and continuing to realize the ultimate development of human-centered UX.
Looking at Jesse James Garrett’s elements of user experience design — “The Elements of User Experience,” including Strategy, Scope, Structure, Skeleton, and Surface, these elements that constitute user experience will likely bring different aspects of change and opportunities with the advent of the AI Agent era. Based on knowledge collection, phenomenon observation, and imagination of future human-machine interaction, I present trend predictions for changes that are likely to transform UX design and web service development after 2025 (estimated 2025–2027) across these five levels:
1. Surface
Surface is the part that users directly interact with, primarily focusing on sensory design. This level integrates multiple sensory experiences including visual, auditory, and tactile, creating pleasant and meaningful experiences through carefully designed interface appearances and interaction feedback. Excellent surface design not only attracts users but also strengthens product brand identity and enhances overall user experience quality.
Trend 1: Boundless Multimodal Interaction
User experience will transcend multiple input method limitations, achieving more natural and fluid human-machine interaction. Users can interact with digital systems through touch, voice, gestures, eye tracking, facial expression input, emotion sensing, and even future brain-computer interaction. The integration of multiple sensory input/output methods will further bridge the gap between humans and systems, bringing digital products or services closer to natural human communication methods and strengthening users’ perceptual memory of brands and products.
2. Skeleton
The skeleton layer deals with more detailed design decisions, including Interface Design, Navigation Design, and Information Design. This level focuses on how functions are concretely presented, how users move through the system, and how information is effectively communicated. Good skeleton design allows users to intuitively understand how to use the product and easily find what they’re looking for.
Trend 2: User Intent (UI)
Future digital services will have the ability to infer and focus on users’ potential intentions, with systems continuously estimating, predicting, and proactively meeting user needs. This transformation shifts traditional interface design from “how to present” Look & Feel themes to “how to anticipate” what users want to do and their possible actions, arranging suitable interaction methods accordingly. This interaction design can reduce users’ cognitive load in function selection, allowing the system to adjust according to human intent, thereby lowering operational difficulty. Through proactive, dynamic UI design, users can focus more on their intended goals rather than operational details.
Trend 3: Node-Based Information Architecture
Information architecture will transform from traditional master-slave, hierarchical relationship structures to more flexible, dynamic, distributed network structures. This revolution makes connections between information more dynamic and organic, allowing users to explore and discover related content more intuitively. Users can explore freely along their own needs or interests, reducing limitations set by predetermined interaction logic information paths.
Trend 4: Hyper-Personalization
Based on intent insights and information node characteristics, interaction experiences will increasingly emphasize extracting more diverse information from users’ behaviors, preferences, and contexts to infer their “current” state and intent, creating highly personalized interaction experiences. This process requires analysis and understanding of large amounts of user data, combined with real-time context awareness, allowing systems to dynamically adjust recommended content, interface configurations, or operational workflows. As a result, the interfaces seen and functional configurations obtained by each user may differ, truly achieving customized, one-to-one interaction effects that vary for each individual user.
3. Structure
The structure layer focuses on system behavior patterns and information organization methods, including two key areas: Interaction Design and Information Architecture. At this level, designers define how users interact with the system and how information is organized and related. Good structural design ensures users feel natural and smooth when interacting with the system while easily finding needed information.
Trend 5: Website as Agent
At the structure layer, websites are no longer just passive functional and service carriers waiting for people to use them or one-way portals for displaying and providing information. Websites will be able to “actively” assist visitors in achieving their goals. When designing websites, UX designers focus more on empathy technology between user intent and website task structures, allowing systems to automatically detect current user needs and handle cross-domain or cross-service complex processes with higher-level computing or planning capabilities. Users can complete goals with fewer operational steps while receiving feedback that aligns with their needs, highlighting the value of websites’ proactive service provision.
Trend 6: Post-Visit Continuity
The relationship between websites and people is no longer limited to “on-site” interactions. Services don’t end when users close pages or applications but continue into other contexts of people’s daily lives. Through “Trend 1: Boundless Multimodal Interaction” as human-machine interaction touchpoints and channels, centered around user intent-driven service purposes, they continue to provide “Trend 4: Hyper-Personalization” to meet people’s needs. Therefore, when designing services, designers must consider long-term interaction continuity and multi/omnichannel experiences both on and off-site. This context-immersive and companionship-style experience emphasizes the deep connection of human-machine interaction, elevating services from fragmented usage to long-term support.
4. Scope
The scope layer transforms abstract concepts from the strategy layer into concrete functional specifications and content requirements. This level defines the boundaries of products or services, clearly stating which functions need to be implemented and what kind of content needs to be provided to meet user needs. Through detailed functional specifications and content requirements, we can ensure product development stays on target, avoiding scope drift or expansion.
Trend 7: Zero-Party Data as Persona
Zero-party data views personal data as a digital extension of user “identity,” meaning that personal data can replace traditional Persona analysis of target groups to more accurately shape individual user characteristics and needs. This concept goes beyond login authorization or single portal integration, allowing users to decide which data can be shared and in what form while maintaining control over their data assets, based on different contexts or needs. This enables services or systems to understand and respond to user intent at a deeper level. As a result, users’ digital footprints become not just a series of anonymized records but a kind of personal “Digital Twin,” allowing various Web or Apps using generative AI to obtain critical behavioral context, providing a more solid data foundation for “Trend 4: Hyper-Personalization” experience design. When constructing this theme, designers face higher requirements for personal data ethics and transparency, ensuring user control over their data and trust in various application scenarios while expanding personalization and understanding depth.
Trend 8: Cross-Platform Information Access
AI Agents act as cross-platform information integrators, becoming bridges between different websites or devices, and making “website users no longer limited to humans themselves” but possibly Agents from other websites or Personal Agents representing specific users. Through human-machine interaction mechanisms authorized to Agents, users only need to communicate goals and intent, effectively reducing steps for repetitive operations like collecting, price comparing, and executing across different websites, making the overall experience more coherent and efficient. However, this situation will also cause some websites to gradually lose the need for human users to visit personally, shifting focus to interaction architectures with AI Agents, forming a new interface design logic that attracts and is friendly to Agent visits.
Trend 9: From UGC to AIGC
Traditional social User Generated Content (UGC) mainly relies on people spontaneously sharing various knowledge, works, and experiences, building the vitality foundation of many platforms, with KOLs and YouTubers deeply participating in this ecosystem through content monetization models. With the advancement of artificial intelligence technology, AI Generated Content (AIGC) can automatically create high-quality text, images, and audio-visual content through language models, and conduct round-the-clock, multi-platform mass distribution when combined with workflow tools such as Zapier, Make, n8n, while also spawning new types of virtual characters that simulate human life sharing. Against this background, humans have begun to distinguish between “human creation” and “machine creation,” and in the online environment, shifted from pursuing “beauty” in evaluation to focusing on seeking “authenticity.” This change reflects the upgrade of human consciousness when facing the rise of AI content, further spurring the demand for deep thinking and critical abilities, also making content providers more cautious in handling content credibility and ethical dimensions.
5. Strategy
The strategy layer is the cornerstone of user experience design, answering the fundamental question of “why” to build this product or service. This level needs to clearly define two core elements: User Needs and Product Objectives. Through deep understanding of users’ real needs and the organization’s business goals, designers can establish the product’s core value proposition and provide clear direction guidance for subsequent design decisions.
Trend 10: AI Sheds New Light on Web3 Adoption
Due to the characteristics of decentralization, wallet as anonymous identity, and smart contract governance, Web3 implements user autonomy over personal data control and transaction permissions under the concept of “zero-party data.” AI Agents allow data to go beyond single-point information and promote the “Data as Service” interaction framework through cross-platform integration capabilities, providing sufficient landing scenarios for Web3. Combined with Web3’s unique token economic model, users can deeply participate in product and community development through tokens, reputation, or other decentralized credentials, elevating from passive “participants” to active “co-creators.” Through bridging between Web2 and blockchain technology, platforms can simultaneously possess the ease of use of traditional networks and the fairness and transparency of decentralization, making the entire ecosystem quickly consolidate long-term and stable user trust and vitality under the dual drive of “human-centered” and “decentralization.”
Trend 11: Collaborative Growth Digital Services
In the AI Agent-dominated internet world, product forms primarily based on knowledge gaps or personalized services, such as online course platforms or knowledge-based internet celebrities, are entering their final stage as AI evolves to provide quality knowledge and situational judgment in real-time. People will begin to realize that traditional roles viewed as “experts” have never truly faced the upcoming rapidly changing and short-cycled turbulent era, thus the “expert myth” no longer holds absolute persuasive power. Instead, communities centered around “trust” and emphasizing “collective growth” to cultivate personal resilience in facing challenges become the core companionship source for people to jointly face future high uncertainty. At this time, genuine interactions and emotional warmth between real people become even more precious. Thus, this community relationship is not just about periodically supplying functions or content but about deep connections where community members nurture each other and grow together in long-term journeys.
We expect to see more service cases emerging around these new human-machine interaction themes in the near future. If you’ve observed other trends in this space, please share your thoughts and insights in the comments below — I’d love to hear your perspective on where you see UX heading in 2025 and beyond.
