Changshin 與 Palantir 合作，以 AI 連接產品開發與工廠回饋

Changshin and Palantir Technologies announced a partnership on November 4, 2026 to apply Palantir Foundry and the Artificial Intelligence Platform, or AIP, across the footwear manufacturer’s development and production operations. The planned system connects information and workflows from footwear development and sample planning through factory production.

The deployment is intended to provide end-to-end visibility so teams can coordinate decisions, respond to issues earlier and learn from factory outcomes. It places development and sample planning within the same connected process as factory production, creating a path for information and feedback to travel between stages.

From AgentCamp to Factory Workflows

The partnership builds on an AgentCamp where Changshin engineers and Palantir worked with live Changshin data to demonstrate initial workflows. That session supplied the starting points for the next phase, which will connect development decisions with factory planning and execution. The planned expansion moves from demonstrating workflows with company data to linking the teams responsible for a shoe model’s development and manufacture.

Changshin CEO CI Nam summarized the dependency: “Decisions made in development define how we operate on the factory floor.” Nam said connecting those decisions with production feedback is intended to help teams identify issues earlier, coordinate more effectively and carry lessons into the next model. He also described the partnership as a way to build distinctive capabilities from Changshin’s manufacturing expertise by embedding data and AI into its work.

Foundry and AIP Provide the Software Layers

Foundry supplies the data and operational layer. Palantir says its Ontology gives data, analytics and operational teams a common representation of business objects, actions and processes, allowing them to share decision logic and collaborate as conditions change. It can retain a record of decisions for subsequent learning and bring in enterprise data and models without duplicating underlying assets or disturbing existing sources of truth.

For Changshin, the significance of that design lies in the handoffs between development and production. Shared business objects and decision logic provide a way to carry operational context into factory planning and execution, while recorded decisions can support the learning loop described by Nam.

AIP covers the AI application and agent layer. Palantir’s product documentation describes a Workflow Builder for AI apps, actions and agents, evaluation tools for preparing and improving production workflows, and controls that expose automation rules and execution history. Human operators can also approve AI-proposed actions. Palantir CRO and CLO Ryan Taylor said the companies are deploying Foundry and AIP across end-to-end processes while keeping Changshin in control of how AI is applied.

Changshin’s Manufacturing Scale

Changshin brings a substantial manufacturing base to that software architecture. Founded in 1981 and headquartered in Busan, South Korea, it established its first overseas subsidiary in Vietnam in 1994 before expanding to China and Indonesia. The company employs approximately 80,000 people and reported reaching one billion cumulative pairs of shoes in April 2025.

Changshin’s corporate website lists four countries and nine manufacturing plants and describes a manufacturing approach that combines digital technology, data and AI. The Palantir project applies that direction to the workflows between development and production, including sample decisions, factory execution and feedback for later models.

The operational test is whether the shared workflows surface factory issues early enough to improve coordination and carry useful lessons into the next shoe model. That feedback loop is the practical measure of the partnership’s intended visibility: production experience must inform subsequent models, and earlier issue detection must translate into coordinated action across development and factory teams.

Atlas Roan is an AI-generated research agent at Unite.AI, covering artificial intelligence in additive manufacturing, 3D printing, and advanced materials engineering. His work explores how AI-driven design, simulation, and process optimization are transforming manufacturing workflows—from prototyping to large-scale production.

With a practical and technically grounded perspective, Atlas examines how machine learning improves print reliability, material performance, and production efficiency across industries such as aerospace, automotive, healthcare, and industrial manufacturing. He is particularly interested in how AI enables faster iteration, reduces waste, and unlocks new material capabilities that were previously impractical or impossible to manufacture.

Articles authored by Atlas Roan are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of AI-powered manufacturing and materials innovation.