# COLMO / 语音交互形象设计 智能体动态设计 - ID: 417 - URL: https://www.stoyard.com/works/417 - English URL: https://www.stoyard.com/en/works/417 - Year: 2025 - Cover: https://static.glabcms.stoyard.com/stoyard/project/cover/20260902/75aqbmnzf96.jpg - Categories: 行业领域：上市公司、先进制造、AI 人工智能, 服务内容：品牌升级（Brand）、视觉创意与传播（Visual & Creative Design）, 技术解决方案：软件设计解决方案 - English Categories: Type of industry field：Listed Company、Advanced Manufacturing、AI Artificial Intelligence, Service Content：Brand Design & Rebranding、Visual & Creative Design, Technical Solutions：Software Design Solution - Author: STOYARD - English Author: STOYARD - Publish: 2025-10-09 - Editor: STOYARD BEIJING ## 中文主体 ### 项目简介 COLMO作为国内高端家电品牌的领导者之一，致力于通过持续的产品创新和科技体验升 级来巩固其领先地位。在生成式人工智能时代，COLMO积极将人工智能大模型技术融入家电家居产品。 COLMO 致力于将生成式人工智能融入高端家居场景，推动智能家电从“被动响应”走向“主动服务”。本项目围绕 COLMO AI 管家，对原有语音交互形象进行系统升级，塑造一个具备多维感知、自然交互与情感表达能力的智能体形象。 通过行业研究、品牌定位梳理与多设备场景分析，我们建立了兼顾识别性、状态辨识度、艺术性、亲密度与 AI 属性的设计体系。最终成果不仅强化了 COLMO 的高端科技形象，也为智能体在不同家电终端中的持续应用与体验扩展奠定了统一基础。 [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260901/246qldgrspm.mp4)     随着生成式人工智能进入家居场景，智能家电的交互方式正在从单一的语音控制，转向具备理解、判断、执行和主动服务能力的 AI Agent。COLMO 发布全维 AI 解决方案 COLMO AI HOME，并推出应用于智能家居场景的 COLMO AI 管家，希望以更加自然、主动且具有情感价值的方式，连接用户、设备与家庭空间。 设计延续 COLMO“高端、极简、科技”的品牌基因，以抽象而富有表意性的视觉语言，重新定义 AI 管家的形象与动态表达。整体造型保持克制、纯粹，通过柔和的形态变化、光影流动和具有节奏感的动态反馈，让无形的语音与智能状态变得可感知、可理解。      [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/fsxg2fq5zqo.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/4k1v6ouw7ix.mp4) “极简”强调功能主义、去繁存简和抽象表达；“科技”则体现在理性秩序、流畅运动、光感材质与智能响应之中。 这些共同构成了 AI 管家的人格基础：不是过度活泼的卡通助手，也不是冰冷、机械的功能图标，而是一位理性、可靠、从容，同时能够给予用户回应与陪伴的智能管家。 [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/oajzpntw8cg.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/wvvief3ywlf.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/t3lvw4dvyy.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/7xu14uytw7u.mp4) 设计将一次完整的语音交互拆解为“默认—预判—唤醒—接收并思考—执行—反馈—恢复”多个阶段，并为每个阶段建立相应的运动和情绪规则。 默认状态下，智能体保持缓慢、平稳的旋转或呼吸感，表现安静而持续的在线状态；当用户靠近或准备发出指令时，运动开始发生轻微变化，形成感知与预判；被唤醒后，转动速度与光线响应增强，明确告诉用户系统已经进入交互状态。 在接收和思考阶段，内部粒子、材质、形态与颜色共同变化，使计算过程获得可视化表达；进入执行与反馈阶段后，动态节奏更加积极，并通过不同反馈方式区分正常完成、警示或异常状态；任务结束后，运动逐渐减速，平滑过渡回默认状态。 [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/16eyet08vj5.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/a8h7kahqkmr.mp4) 2025年7月28日，AI科技家电品牌COLMO在2025世界人工智能大会开放魔盒直播间重磅发布其全维AI解决方案——COLMO AI HOME，并呈现了智能家居领域首个已实现应用的AI Agent智能体——COLMO AI 管家，作为行业首个定位AI的高端家电品牌， COLMO展示出引领智能家居迈向主动服务的技术实力。 COLMO AI 管家需进入空调、冰箱、厨电以及其他不同尺寸的家电终端，因此设计从一开始就考虑了跨设备的一致性。  [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260909/oe2udcdsk1e.mp4) 在项目初期，我们对市场上大量的智能体形象进行了系统性调研与分析，从形象特征、交互方式、情感表达等多个维度进行归纳与对比。通过研究，我们将当前主流的智能体形象大致分为三类：  通过对不同设备形态和交互载体的分析，我们发现：屏幕尺寸与智能体形象类型之间存在显著的相关性。 在大屏设备中，品牌更倾向于采用拟人化的智能体形象，以增强情感连接与沉浸式互动体验；而在小屏或无屏设备中，出于信息承载与交互效率的考虑，更多选择抽象型甚至传统型的形象设计，以保持简洁与功能聚焦。 目前，家电行业在智能体形象的发展仍处于初中期阶段。整体设计的差异性相对有限，品牌识别度不高，缺乏鲜明的视觉与情感特征。各品牌正处于探索期，尝试在理性科技与情感陪伴之间寻找平衡，以建立更符合家电场景特性的智能体定位与视觉语言体系。     我们首先对 COLMO 品牌的核心定位进行了系统梳理与提炼。 品牌关键词为：高端（Premium）、极简（Minimalist）、科技（Technological）。围绕这三大关键词，我们展开了多轮的发散与收敛分析，探索品牌在智能时代应具备的视觉气质与人格特征。 这一阶段确立了设计的总体方向——在高端理性与科技秩序之上，注入克制、纯粹且富有温度的设计表达。   在品牌方向确立后，我们建立了智能体视觉特征的五维雷达图模型，五个核心维度包括：识别性、状态辨识度、设计感与艺术性、亲密度、以及AI属性。通过雷达图的可视化分析，我们清晰地评估了现有形象在各个维度上的优势与不足，为后续设计策略提供了量化依据。  进一步地，我们从行业维度构建了智能体的形象坐标轴模型，通过对市场主流智能体的分布分析，我们将 COLMO 的智能体定位在坐标轴的中心位置，略偏向“抽象”与“表意”方向。  在本次智能体设计过程中，我们面临了多重交互层面的设计挑战：情绪表达的克制与明确之间的平衡，状态切换的流畅与连贯性，多设备场景下的视觉一致性。  [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20251009/ont3phgthbp.mp4) 在交互层面，我们重点处理唤醒、聆听、思考、回应及情绪反馈等状态，使不同阶段之间的切换更加自然连贯；同时控制情感表达的尺度，在科技理性与陪伴温度之间取得平衡。针对不同尺寸与形态的家电终端，设计采用可扩展的视觉规则，确保智能体在多设备环境中保持统一、清晰且具有品牌辨识度的体验。 未来，智能体的视觉定义不再是依附于某一块屏幕的单独动画，而成为能够贯穿 COLMO 智能家居生态的统一视觉角色。  ## English Content ### Introduction As one of the leaders of domestic high-end home appliance brands, COLMO is committed to consolidating its leading position through continuous product innovation and technological experience upgrading. In the era of generative artificial intelligence, COLMO actively integrates artificial intelligence large model technology into home appliances. COLMO is committed to integrating generative artificial intelligence into high-end home scenarios, and promoting smart home appliances from "passive response" to "active service". This project revolves around the COLMO AI steward, systematically upgrades the original voice interaction image, and creates an intelligent body image with multi-dimensional perception, natural interaction and emotional expression capabilities. Through industry research, brand positioning combing and multi-device scenario analysis, we have established a design system that takes into account identification, state identification, artistry, intimacy and AI attributes. The final result not only strengthens the high-end technology image of COLMO, but also lays a unified foundation for the continuous application and experience expansion of the agent in different home appliance terminals. [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260901/246qldgrspm.mp4)          [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/fsxg2fq5zqo.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/4k1v6ouw7ix.mp4) "Minimalism" emphasizes functionalism, de-complexity and simplicity, and abstract expression. "Technology" is reflected in rational order, smooth movement, light-sensing materials and intelligent response. These together constitute the personality basis of AI housekeeper: it is not an excessively lively cartoon assistant, nor a cold and mechanical function icon, but an intelligent housekeeper who is rational, reliable and calm, and can give users response and companionship at the same time. [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/oajzpntw8cg.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/wvvief3ywlf.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/t3lvw4dvyy.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/7xu14uytw7u.mp4) The design disassembled a complete voice interaction into multiple stages of "default-anticipation-awakening-receiving and thinking-execution-feedback-recovery", and established corresponding movement and emotional rules for each stage. In the default state, the intelligent body maintains a slow and smooth rotation or breathing sense, showing a quiet and continuous online state. When the user approaches or prepares to give instructions, the movement begins to change slightly, forming perception and prediction. After being awakened, the rotation speed and light response are enhanced, clearly telling the user that the system has entered the interactive state. In the receiving and thinking stage, the internal particles, materials, shapes and colors change together, so that the calculation process can be visually expressed. After entering the execution and feedback stage, the dynamic rhythm is more active, and normal completion, warning or abnormal states are distinguished through different feedback methods. After the task is over, the movement gradually slows down and smoothly transitions back to the default state. [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/16eyet08vj5.mp4) [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260910/a8h7kahqkmr.mp4) On July 28, 2025, AI technology household appliance brand COLMO released its full-dimensional AI solution -- COLMO AI HOME in the open magic box studio of the 2025 world artificial intelligence conference, and presented the first AI Agent intelligent body that has been applied in the field of smart home -- COLMO AI butler. as the industry's first high-end household appliance brand with AI positioning, COLMO demonstrated its technical strength to lead smart home towards active service.  [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20260909/oe2udcdsk1e.mp4) At the beginning of the project, we conducted systematic research and analysis of a large number of intelligent body images in the market, and summarized and compared them from multiple dimensions such as image characteristics, interaction methods, and emotional expression. Through our research, we broadly divide the current mainstream intelligent body images into three categories:  Through the analysis of different device forms and interaction carriers, we find that there is a significant correlation between screen size and intelligent body image type. In large-screen devices, brands are more inclined to adopt anthropomorphic intelligent body images to enhance emotional connection and immersive interactive experience, while in small-screen or non-screen devices, for the sake of information carrying and interaction efficiency, more abstract or even traditional image design is chosen to maintain simplicity and functional focus. At present, the home appliance industry is still in the middle stage of the development of intelligent body image. The difference of the overall design is relatively limited, the brand recognition is not high, and the distinctive visual and emotional characteristics are lacking. Brands are in an exploratory period, trying to find a balance between rational technology and emotional companionship in order to establish an intelligent positioning and visual language system that is more in line with the characteristics of home appliance scenes.        Furthermore, we have constructed an image‑coordinate‑axis model for AI agents from an industry‑level perspective. By analyzing the distribution of mainstream AI agents in the market, we have positioned COLMO’s agent at the center of the axis, with a slight bias toward the “abstract” and “expressive” quadrants.  In the process of intelligent body design, we are faced with the design challenges of multiple interaction levels: the balance between restraint and clarity of emotional expression, the smoothness and consistency of state switching, and the visual consistency in multi-device scenarios.  [Video](https://static.glabcms.stoyard.com/stoyard/gallery/20251009/ont3phgthbp.mp4) 