Xsight Group Featured in Harvard Business Review for AI Agent Orchestration Across Business Silos

Hong Kong-headquartered cross-border technology group Xsight has been featured as a research case in Harvard Business Review’s September-October 2026 issue, in an article examining how AI agents can orchestrate work across organisational silos.

The article, titled “How AI Agents Orchestrate Work Across Silos”, examines how a new generation of AI systems can connect business functions and reduce the coordination friction that can slow enterprise decision-making. Xsight was cited for its approach of breaking complex cross-departmental decisions into modular tasks and using a layered agentic AI orchestration system.

The Harvard Business Review report points to the challenge faced by companies when decisions involve multiple functions including procurement, logistics, finance, legal, marketing, sales and merchandising. While individual AI tools can support specific tasks, the article notes that independently operating AI agents can produce conflicting outputs when there is no system coordinating their work.

Xsight's approach is based on dividing complex decisions into precise and defined tasks, with each task assigned specific inputs, outputs, constraints and objectives. The company also standardises information and assumptions so that different functions work from common definitions.

Building on this structure, Xsight has developed a multi-layered agentic AI orchestration system. At the task level, specialised AI agents handle functions including contract analysis, demand forecasting and capacity planning.

At the orchestration level, connectors check that outputs remain within management-defined parameters and route validated information between workflows. A master AI orchestrator then structures and coordinates the analysis and human input required for broader enterprise decisions.

Xsight founder and CEO Lewis Chan said the system was not intended to replace management in making final business decisions, but to provide an operational backbone for task distribution, information synchronisation and conflict validation.

The objective, he said, was to reduce the time managers spend on alignment meetings and allow them to concentrate on higher-level commercial judgement.

The company's appearance as a Harvard Business Review research case comes as Xsight continues to develop its cross-border supply chain technology business.

Its core platform, UX168, is described as a data-driven cross-border supply chain e-commerce platform serving more than 80 million end consumers globally through more than 30 international sales channels. The group has research and development, operations and supply chain centres across Hong Kong, Shenzhen, Guangzhou, Huizhou, the United States, the Netherlands and Japan.

Xsight has expanded its supply chain infrastructure over the past two decades, including a US overseas warehouse established in 2013, entry into Japan in 2017 and a Netherlands fulfilment centre launched in 2024.

In 2025, the company operationalised a digital origin warehouse in Zhongshan and a 120,000-square-metre intelligent mega-warehouse in Huizhou, integrating technologies including the Internet of Things, intelligent sorting and predictive data forecasting.

At the centre of its operations is Xsight's proprietary XsOS operating system, which standardises and modularises processes across the business value chain, covering product selection, sampling, pricing, fulfilment and marketing.

According to the company, the system has helped double its new-product hit rate from 10% to more than 20%, while supporting compound annual growth of more than 40%.

Xsight said the rapid development of large language models is also prompting a strategic shift, with the group seeking to move from a technology-enabled trading enterprise towards an open, vertical-industry digital supply chain platform.

The development reflects a broader shift in enterprise AI from standalone applications towards systems designed to coordinate information, workflows and decision-making across organisational boundaries.