How Enterprises Can Harness Agents? They Provided Their Answers at WAIC

On July 19, during the 2026 World Artificial Intelligence Conference (WAIC), a special forum titled "Mastering Enterprise-Level Agents" hosted by DingTalk was held at the Shanghai Expo Center. Focusing on the implementation roadmap, organizational practices, and future directions of enterprise-level agents, the event brought together hundreds of AI experts and industry pioneers to exchange insights on the new productivity that agents bring to enterprises.

Guests including Zhu Hong, Head of Core Platform Business at DingTalk; Shri Narayanan, Chair Professor at the University of Southern California and Member of the National Academy of Engineering; Gu Chao, Chief Security Officer at STO Express; Xu Jiaxin, General Manager at Kootai; Li Ren, Chairman of Zhongyi Group and Founder of Opunna; Yi Zihan, Founder and Chairman of Songxianxian; Cong Zelong, Director of AI Product Center at Pisen; and Wei Jun, Chairman of Zhejiang Ucola Intelligent Technology, attended and shared their valuable perspectives.

From Pilots to Scale: The Evolution Path of Enterprise-Level Agents

Looking back over the past year, nearly every company has been discussing, applying, and developing agents. Yet how to truly integrate agents into core operations—transforming them from experimental tools into stable, productive systems and scaling them beyond isolated pilots—remains an unsolved challenge without a standardized model.

According to Zhu Hong, the key challenge for enterprises is no longer whether they can build one agent, but whether they can mass-produce, operate, and manage large numbers of agents. Since each agent carries marginal costs, businesses must extract tangible value from them. He predicts that in the future, every enterprise may employ numerous digital workers, requiring not just standalone tools, but an entire infrastructure that supports secure collaboration, continuous operation, and ongoing evolution of agents.

Building an Agent Collaboration Network: Rethinking DingTalk’s Role in the AI Era

In response to this trend, DingTalk is redefining its role. Zhu Hong admitted that over the past decade, DingTalk primarily connected people with organizations. In the future, it aims to connect people with agents, agents with agents, and agents with organizations—building a collaborative network tailored for production-grade agents.

This is DingTalk's vision for the next ten years—an AI platform friendly to agents, developers, and corporate data assets. To achieve this, DingTalk is building an AI-powered workplace platform to help individuals own personalized AI assistants, while simultaneously advancing the AI DingTalk ecosystem to enable enterprises to build, coordinate, and operate digital employees more efficiently.

"Tools are merely the starting point; practice is what truly matters—that's exactly why we're holding this forum today," said Zhu Hong. He expressed his hope that this new AI-powered productivity platform could walk side by side with more enterprises and successfully answer the critical question: How do we master the use of agents?

Industry Cases: Integrating Agents into Core Business Workflows

During the enterprise AI practice session, representatives from logistics, consumer retail, manufacturing, and intelligent technology sectors shared frontline experiences in deploying enterprise-level agents.

STO Express, one of China’s largest economy courier service providers operating under a franchise model with approximately 350,000 employees, manages both physical production safety and online information security. Chief Security Officer Gu Chao pointed out that AI is evolving from a tool for efficiency enhancement into a vital capability for risk governance.

For example, in hazard identification, photos taken by staff might contain multiple safety risks, which are difficult for a single safety officer to fully detect. Today, AI leverages self-trained skills and knowledge bases to automatically identify risks, propose solutions, and through DingTalk, notify responsible personnel and track rectification progress—achieving a complete closed-loop process from detection to resolution.

Kootai, an emerging retail brand offering whole-home storage solutions characterized by high order values, high service demands, and strong experiential focus, operates around 200 service provider stores. General Manager Xu Jiaxin emphasized that the true value of AI lies in whether it genuinely integrates into business processes.

She cited customer content acquisition as an example: previously, social media operations relied heavily on individual capabilities and yielded limited output. Now, AI handles breaking down viral content, generating copy and scripts, mixing visual materials, and reviewing livestream performances. Frontline teams also use AI for automated customer service replies and private domain outreach, significantly improving the efficiency and consistency of content production. Since deeply integrating AI into workflows in March this year, within just four months, every employee’s work interface has evolved into “AI on the left, work on the right.”

Pisen was among the first companies to invest in AI hardware, launching the world’s first desktop AI robot at CES in January. Cong Zelong, Director of the AI Product Center, stressed that if engineering expertise remains confined to individuals, it becomes a bottleneck for R&D.

To address this, Pisen connected its decade-long accumulated R&D system and proprietary component database to its AI platform. After inputting product requirements, agents can automatically extract parameters, search for components, calculate BOM costs instantly, find alternative solutions when budgets are exceeded, and conduct adversarial reviews to fill gaps. The system ultimately outputs initial drafts of hardware topology diagrams and specification sheets, shortening a development cycle that used to take at least one month to under a week. The team also built a Skill-sharing platform, transforming individual experience into reusable assets. So far, over 50 high-quality Skills have been accumulated, with more than 600 code submissions.

Meanwhile, Li Ren, Chairman of Zhongyi Group; Yi Zihan, Founder and Chairman of Songxianxian; and Wei Jun, Chairman of Zhejiang Ucola Intelligent Technology, also shared their practical applications based on DingTalk AI. These cross-industry cases demonstrate how enterprise-level agents are moving from pilot projects into core business operations, highlighting the profound value of DingTalk AI in cost reduction, efficiency improvement, risk control, and organizational collaboration.

The Essence of AI Implementation: Iterating Organizational Effectiveness

During the conference, DingTalk showcased its latest achievements in enterprise-level agents and AI collaboration at Alibaba Group’s exhibition area. Products such as AI Voice Recorder, AI Notetaker, AI Spreadsheet, and the AI Productivity Platform attracted many visitors who stopped to experience features covering meeting minutes, document processing, PPT creation, information search, data coordination, and workflow tracking.

The enthusiastic response validated an important trend: implementing enterprise-level agents is not merely a technical issue—it involves organizational collaboration, process redesign, knowledge accumulation, and the sustained ability to operate AI systems. The real value of agents does not lie in demonstrations or isolated tools, but in actual business impact and embedded organizational capabilities.

We dedicated to serving clients with professional DingTalk solutions. If you'd like to learn more about DingTalk platform applications, feel free to contact our online customer service or email at This email address is being protected from spambots. You need JavaScript enabled to view it.. With a skilled development and operations team and extensive market experience, we’re ready to deliver expert DingTalk services and solutions tailored to your needs!