Zhaogang.com Appears at 2026 Global Data Business Conference: Driving Industrial AI Implementation with Real Transaction Data

Lead:

From October 30 to 31, the 2026 Global Data Ecosystem Conference, themed "Symbiosis of Data and Intelligence, Connecting the World," was held at West Bund in Xuhui, Shanghai. The conference focused on data element circulation, international digital cooperation, and data-intelligence technology innovation, building a platform for exchange and discussion, industry matchmaking, and showcasing achievements.

On October 31, the data frontier forum "Building a Data Foundation, Winning the Future with Quality—High-Quality Datasets Empowering AI Innovation and Development" was held, hosted by Shanghai Data Group, Shanghai Data Research Institute, and other organizations. Zhang Xiaokun, Managing Partner and Vice President of Zhaogang.com Group, was invited to deliver a keynote speech titled "High-Quality Datasets Driving Industrial Development," sharing Zhaogang.com's practice of building high-quality industry datasets from real transaction data to support the R&D and application of industrial intelligent agents, and participated in the "Joint Construction of Dataset Open Source Community" ceremony on behalf of the group.

As large model capabilities continue to improve, how AI can truly understand industries and enter enterprise processes has become an important topic in industrial applications. In his speech, Zhang Xiaokun pointed out that the capabilities of vertical industry intelligent agents depend not only on general-purpose models, but also on whether high-quality industry data can be obtained and whether such data can support actual business judgment and execution.

B2B trade is a typical "conversation-driven" business. From inquiries and quotations to sourcing goods, and on to contracts, order placement, and fulfillment, a large amount of information is generated in continuous communication, including industry abbreviations and specification conventions as well as constantly changing prices, inventory, and delivery conditions.

In 14 years of industrial practice, Zhaogang.com has continuously accumulated real transaction data, customer communication data, fulfillment service data, and mid-platform operational processes, providing support for demand understanding, resource matching, and business collaboration. Currently, Zhaogang.com's AI intelligent agents execute more than 600,000 tasks per month, improving overall work efficiency by more than 60%.

Not just "feeding corpus," but also accumulating replicable organizational capabilities

Having a large amount of data does not mean having usable high-quality datasets.

B2B enterprises' quotations, documents, and customer information are often scattered across chats, files, business systems, and personal experience. If information is not aggregated, AI lacks the context needed to understand the business; and with only fragmented tools, it is difficult to form complete business collaboration.

Zhaogang.com integrates its industry accumulation into Skills, MCP, and Agentic workflows, connecting enterprise data, business systems, and Agent execution, so that experience moves from "staying in people's heads" to becoming replicable organizational capability.

Take "Zhaogang Lobster," a transaction-oriented AI Agent for the steel trading industry, as an example. The product connects demand identification during communication with tasks such as quotation, CRM updates, business system operations, and to-do reminders, promoting the integration of information flow and trade flow.

High-quality datasets ultimately need to demonstrate value in business.

The relevant capabilities are exported to enterprises through products such as Zhaogang Lobster and NextB2B. Among them, NextB2B targets the B2B trade circulation sector, centering on four types of AI experts—sales, procurement, back-office, and operations—and covering areas such as customer acquisition and quotation, sourcing and price comparison, order entry and reconciliation, and business analysis. Zhaogang.com is extending the same set of AI application methods to fields such as electronic components, electrical and electronics, chemicals and plastics, freight forwarding and logistics, and non-ferrous metals, exploring a leapfrog development path for "conversation-driven" B2B enterprises.

Participating in joint community construction, exploring synergy between data and industrial applications

At the forum, Zhang Xiaokun, on behalf of Zhaogang.com Group, joined representatives of Kupas, Shanghai Data Group, and other organizations in the "Joint Construction of Dataset Open Source Community" ceremony.

To connect the data supply and demand chain and promote the efficient circulation of data elements, Shanghai has built the OpenDataHub dataset open source community, which officially launched on July 17 this year. It brings together high-quality multimodal datasets from 10 industries, currently lists more than 600 datasets, and is open to AI enterprises, research institutions, and government departments. Building high-quality datasets requires not only technical and research strength, but also real problems, application experience, and validation scenarios from the industrial front lines. Participating in the forum and the joint community construction is also a practice by the group in carrying out exchange and cooperation around high-quality dataset construction and industrial AI applications.

In the future, Zhaogang.com Group will continue to improve the evolutionary closed loop of "real industry datasets—AI intelligent agent teams—full-chain industrial AI services," promote the release of data value on the basis of security and compliance, and transform industry accumulation into AI capabilities that enterprises can use.