KunYan Launches Jewelry Agent Globally for the First Time at World AI Conference

Recently, at the WAIC 2026 Jewelry Design themed event, KunYan AI publicly released its first industry intelligent agent for jewelry new product development — the KunYan Jewelry Agent. The product is built on KunYan AI's self-developed jewelry generation large model as the core foundation, connecting key links in the entire jewelry new product development process, including creative design, 3D modeling, craft analysis, cost management, marketing content, and supply chain collaboration.

As one of the earliest enterprises engaged in jewelry vertical large model and Agent industrial practice in China, KunYan AI released the KunYan Jewelry Design Large Model in December 2023. In June 2025, KunYan AI launched the internal testing version of the Jewelry Design Agent; on July 19, 2026, the KunYan Jewelry Design Agent was officially released at the WAIC 2026 Jewelry Forum. This launch extends KunYan's jewelry AI capabilities from single-point content generation to product understanding, task planning, professional tool invocation, version management, and cross-link collaboration.

Liu Shuchun, Founder and CEO of KunYan AI, PhD in AI from Fudan University, delivered a keynote speech titled "AI from Tool to Partner: Reconstructing the Underlying Logic of the Jewelry Industry Chain," systematically sharing practical insights on AI entering the jewelry new product development process, covering topics such as jewelry generation large models, unified product data, 3D structured modeling, craft constraints, cost analysis, Agent workflows, and human-machine collaboration boundaries.

In recent years, generative AI has significantly improved the production efficiency of jewelry creative proposals. However, transforming a design rendering into a prototype-ready, manufacturable, and sellable product still requires multiple stages: brand positioning adaptation, structural feasibility confirmation, 3D modeling, material selection, gemstone setting process judgment, cost evaluation, sample modification, commercial content production, and market feedback. After rendering efficiency improvements, information transfer losses, scattered versions, repeated communication, and cross-link waiting have become new efficiency bottlenecks.

news.waic.para05

In the concurrent roundtable discussion, Liu Shuchun stated that the core competitiveness of vertical industry AI is not limited to drawing capabilities, but lies more in the understanding of professional knowledge, craft logic, and actual implementation conditions. The long-term value of jewelry AI is to establish more stable connections between creative generation and 3D, craftsmanship, cost, prototyping, content, and supply chain.

The KunYan Jewelry Agent is not a simple assembly of multiple AI functions. Its core technical foundation is the KunYan Jewelry Design Large Model independently developed by KunYan AI. The model is continuously trained and iterated for jewelry, accessories, and related craft scenarios, with emphasis on understanding and generating capabilities for shape contours, material gemstones, style semantics, brand language, structural relationships, and wearing scenarios.

Driven by jewelry industry needs, KunYan AI has accumulated and refined over 1 million high-definition jewelry images and 200,000 high-precision 3D data resources, and combined with design proposals, hand-drawn sketches, craft feedback, prototyping records, and marketing materials, continuously improving the industry data and knowledge system covering "image—structure—craft—product." Compared to general image generation models, the KunYan Jewelry Generation Large Model focuses more on jewelry professional details, brand expression, and the product materialization process.

Building on the large model capabilities, the KunYan Jewelry Agent further integrates context understanding, long-term project memory, missing information completion, multi-step task planning, professional tool invocation, and feedback verification capabilities. It is not just a one-time content generation tool, but an intelligent collaboration system that continuously works around real new product goals.

A new jewelry product is not a static design rendering, but a continuously evolving product digital object, carrying design intent, material gemstone parameters, structural dimensions, 3D versions, craft constraints, cost information, commercial content, prototyping records, and market feedback. The KunYan Jewelry Agent uses unified product data as the core, enabling design, craft, supply chain, and marketing roles to collaborate around the same product version.

Liu Shuchun stated that the Agent's goal is not to replace humans in making all decisions, but to reduce information omissions, waiting costs, and repeated communication in the process. Brand expression, cultural element usage, critical structural safety, mass production commitments, copyright compliance, and commercial responsibilities are still ultimately confirmed by designers, craftsmen, and enterprise leaders.

Liu Shuchun, Founder and CEO of KunYan AI, holds a PhD in Artificial Intelligence from Fudan University, with over a decade of experience in AI algorithm R&D, large-scale system construction, and industrial application, focusing on computer vision, natural language processing, multimodal large models, generative AI, and Agent systems.

Prior to founding KunYan AI, Liu Shuchun served as a Senior Algorithm Expert at Alibaba, NLP Algorithm Lead at Qiniu Cloud, and Algorithm Director at Boyin Hechuang, participating in the R&D and industrial application of multiple large-scale internet AI systems and multimodal understanding and generation algorithms. His team won the COCO 2020 Panoptic Segmentation Competition championship, and Liu Shuchun is also one of the authors of AI technology books including "Python Natural Language Processing in Practice."

From frontier algorithm research, large-scale internet system construction, to jewelry vertical AIGC entrepreneurship, Liu Shuchun continues to focus on the same question: how to bring AI into real industry scenarios, transforming model technology into products and processes that are implementable, verifiable, and capable of continuously generating business value.

KunYan AI has not limited its enterprise AI capabilities to a single model, but has progressively built an enterprise-level AI capability system covering data, knowledge, models, tools, Agents, and governance on the foundation of the jewelry design large model.

In this system, the KunYan Jewelry Generation Large Model is responsible for understanding design intent, product semantics, and brand needs, and generating professional content; professional tools such as image, 3D, CAD, rendering, and enterprise business systems handle specific task execution; the KunYan Jewelry Agent undertakes goal decomposition, task planning, information completion, tool invocation, version management, and cross-link collaboration; design direction, aesthetic choices, critical craft judgments, and final decisions and responsibilities remain with humans.

At the industry infrastructure level, KunYan AI is working with jewelry associations to advance AIGC-related standard construction for jewelry, and participating in jewelry 3D design and manufacturing standardization projects, promoting the gradual structuring of generation processes, quality evaluation, manufacturing constraints, and responsibility boundaries.

In the future, KunYan AI will continue to deepen R&D and industry practice around jewelry generation large models, jewelry design Agents, 2D-to-3D structuring, editable 3D models, craft constraint checking, digital asset management, and multi-Agent collaborative workflows, driving AI from creative tools further into the core of jewelry new product development.

KunYan AI is a technology enterprise focusing on AIGC applications in the jewelry and accessories industry, conducting product R&D and industry practice in AI jewelry design, jewelry generation large models, Jewelry Agent, image-to-3D, digital asset management, marketing content generation, enterprise model customization, and design-manufacturing collaboration, dedicated to making AI serve jewelry brands, design studios, and manufacturing enterprises more deeply.