DingTalk "grows downward," rooting AI in every business place

This article is from WeChat Official Account | Choutai Author | Shi Can

The future of AI applications isn't only in Beijing, Shanghai, Guangzhou, and Shenzhen

In March, the air across China was filled with a sense of "AI restlessness." DeepSeek went viral. Social media feeds were flooded with generative content. Business owners' anxieties rapidly intensified: If you don't understand AI, will you be left behind by the times? After the Spring Festival, conferences, forums, and training sessions sprouted up like mushrooms after rain in Hangzhou, Zhejiang—so long as AI was mentioned, it seemed to point toward the future. In this atmosphere, Wu Che's schedule quickly filled up.

Wu Che was originally part of DingTalk’s early startup team and now serves as Chief AI Officer at Xinfengwei, a digital service provider that grew within the DingTalk ecosystem. Previously, his work rhythm was steady and manageable—clients were clear, needs defined, paths predictable. He helped companies digitize their organizations, improve workforce efficiency, and reduce cloud costs. It was an orderly kind of busyness.

Since the global explosion of the U.S.-developed AI product ChatGPT in November 2022, Chinese enterprises have experienced several waves of AI anxiety. At first, most business owners were merely surprised and curious, with only mild unease. By spring and summer of 2023, anxiety deepened into experimentation—but practical use cases remained blurry and methods unclear. In 2024, structural anxiety emerged: strategic levels within companies had direction but couldn’t find capable implementers.

After March 2025, everything changed. Domestic Chinese AI models exploded in capability, and corporate demands suddenly became both vague and urgent—"We must do AI"—yet no one could clearly explain what exactly "AI" could do.

Industrial anxiety drives industrial demand, making Wu Che a “lifeline” for many. He constantly receives calls from company leaders across different industries, inviting him directly into phone meetings, offline events, and industry summits. "The biggest difference this year is that clients are proactively reaching out," he said. "But that initiative feels more like collective confusion."

This anxiety and confusion aren’t limited to a few isolated businesses. At a Dingsummit held in Hefei, Anhui, over three hundred pairs of eyes gathered together—the urgency and uncertainty spreading visibly through the venue.

October 30, on-site at the DingTalk Anhui Summit, crowds bustling

Since the beginning of this year, Hefei has become one of the fastest-growing cities for DingTalk's business. After the launch of AI-powered DingTalk 8.0, many companies actively requested access to AI features. Across Anhui province, more than 120,000 enterprises now use DingTalk, with 50,000 located in Hefei—about 40%. Among Hefei’s 10,000-plus large-scale enterprises, roughly half have already integrated DingTalk.

The surge in Anhui wasn’t accidental. On one hand, Hefei is known as the "city of venture capital," where emerging industries—quantum technology, photovoltaic new energy, and power batteries—have brought advanced productivity and management awareness. On the other hand, local governments introduced a series of digitalization and intelligent transformation policies, helping companies realize digital transformation is no longer optional but essential. Zhang Ziyou, regional head of DingTalk in Anhui, works to turn awareness into action through summits, training, and enterprise services.

Yet, a huge gap remains between ideals and reality. Shaking his head with a smile during a video call, Wu Che said: "On short-video apps, we’re bombarded with AI every day—it feels like AI already rules the world. But when you actually step into real companies, you’ll find many still can’t even properly use basic office software."

A research report released in May 2025 showed that Chinese companies’ investment in AI is rapidly increasing, nearly doubling year-on-year, with over 40% of funds going toward generative AI. Among nearly 3,000 surveyed enterprises, about half have started deploying AI, though most remain in the early stages of application.

Even more complex are regional disparities and knowledge gaps. In 2024, the eastern region saw a 6.5% year-on-year growth in digital industry revenue, accounting for 73.6% of the national total, while central, western, and northeastern regions grew only 4.2%, 0.8%, and 2.5% respectively. Digital development remains concentrated along coastal areas, with inland regions lagging significantly in infrastructure, talent, and awareness. According to Wu Che’s observations, southern city enterprises tend to be more pragmatic, focusing on tangible AI implementation, whereas northern counterparts mostly remain at the stage of “understanding trends.”

To make AI integrate more efficiently into industries, DingTalk chose to take proactive steps. Within less than two months, DingTalk’s “AI Hundred Cities Initiative” had covered 11 cities nationwide, hosting 11 Dingsummits and attracting over 1,000 participating enterprises.

Within DingTalk, there’s a shared belief: the future of AI applications lies not only in Beijing, Shanghai, Guangzhou, and Shenzhen, but also in so-called “new first-tier” cities and industrial hubs like Ningbo, Hefei, Wuxi, and Jinan.

This job can’t float in the sky—it must stay grounded

Yiwu, a city famous for commerce, hosts 1.2 million registered business entities, of which 860,000 are individual households. Its economic structure is highly fragmented yet exceptionally vibrant. Almost everyone here talks business and does business.

In October, Mo Shang traveled south from Hangzhou to Yiwu, visiting businesses, conducting surveys, monitoring online activity, identifying latent needs among small and medium-sized enterprises (SMEs), and co-designing customized solutions.

Prior to this, he mainly served mid-to-large-sized companies. In Yiwu, however, most enterprises employ fewer than ten people. These companies typically lack formal “management departments,” and growth matters far more than processes.

So Mo Shang adjusted his promotion strategy.

"Before, we promoted HR, finance, and administration functions. Now they care more about whether it saves money and earns profits," he said.

For e-commerce clients, he recommends AI spreadsheets. Merchants used to hire designers to create images, producing four or five per day at high cost. Now, AI fields within DingTalk’s AI spreadsheet can generate them in bulk.

"Copywriting, translation, inventory tracking—all possible. We help them build a CRM system using AI spreadsheets." Such features are down-to-earth, practical, direct—and perfectly aligned with Yiwu traders’ style. "Tell them it saves money and increases traffic, and they’re willing to adopt it."

But Yiwu’s businesses don’t exist in isolation. Mo Shang discovered that despite their small size, these firms are embedded in complete supply chains. Behind each merchant often lie contract manufacturers linked to manufacturing, warehousing, logistics, and foreign trade. "Talk to a shopkeeper—he doesn’t own a factory himself, but several factories produce for him. Dig deeper, and you connect straight to the manufacturing side. The deeper you go, the wider it gets."

This structure forced Mo Shang to develop a “layer-by-layer infiltration” path for DingTalk’s implementation. He believes future AI competition won’t just be about algorithms, but also computing power and data. While computing power forms the foundation, data is the lifeblood—and its source lies precisely in tens of millions of SMEs across the country.

"What DingTalk is doing is essentially building pipelines," Mo Shang likens it to plumbing. "Only when we lay these pipes properly and deliver services to perfection can water, electricity, gas—that is, AI capabilities, computing power, and data—truly flow into every small business."

He paused, then added, "This job can’t float in the sky. It must stay grounded—even dig beneath the surface."

As Mo Shang rooted himself in Yiwu, the Yiwu–Ningxia Chamber of Commerce also became a key node in this “pipeline construction project.” The chamber gathers over a hundred member companies involved in translation and foreign trade, having witnessed every phase of Yiwu’s industrial evolution. Secretary-General He Yinglong has used DingTalk since 2015 and is now leading the chamber in experimenting with AI to reshape enterprise management.

With support from DingTalk’s service team, the chamber began systematically introducing AI spreadsheets, smart approval functions, and more, helping members cut labor costs and build basic digital systems. Some foreign trade companies adopted AI translation and image generation, improving overseas communication and product presentation. Small manufacturers replaced outdated traditional software with AI spreadsheets for inventory management.

Yiwu is like this; so too is Anhui, hundreds of kilometers away.

As DingTalk’s regional lead in Anhui, Zhang Ziyou has crisscrossed nearly the entire province over the past six months. Whether in manufacturing parks in Wuhu or innovation incubators in Hefei, whenever someone mentions “DingTalk AI,” business owners immediately crowd around asking, “Can you tell us how to get started?”

On October 30, the Anhui Dingsummit offered 300 spots—registration filled up in under a day. An hour before the event started, people were already scrambling for seats.

In Anhui, AI adoption feels more like an “awakening movement”—awareness determines corporate action and shapes strategy. Zhang Ziyou leads his team, partners, and government collaborators to push entrepreneurs from “watching the spectacle” to “making real investments.” Policies, industries, and atmosphere in Anhui provide fertile ground for this shift.

Take Laoxiangji, for example—a chain with over a thousand stores nationwide, requiring massive employee training. After adopting AI-based training evaluation, new staff training cycles shortened by one-third, with efficiency multiplied several times. Or Sungrow Power Supply, which built its own AI assistant platform. With 27 intelligent assistants operating across departments, administrative, financial, and operational workflows all accelerated.

Yiwu and Anhui—two seemingly different regions—represent two typical models of DingTalk’s AI implementation: one extends pipelines into the densest industrial fabric; the other ignites change in the most dynamic manufacturing bases.

Together, they point to one trend—AI revolution will ultimately happen only where people open shops and do business.

'DingTalk Photo Documentation Method' in State-Owned Enterprises

Digital transformation differs between private and state-owned enterprises. Private firms prioritize innovation and efficiency, while SOEs place data security and domestic deployment first. Achieving digital innovation in SOEs is much harder.

Nonetheless, innovation still emerges—and grows—within SOEs.

Li Zhengjun, a technical officer at a state-owned enterprise, decided to use technology to solve port safety issues. What gave him confidence wasn’t just technical expertise, but years of accumulated operational experience—and tools like DingTalk and Tongyi Qianwen.

Over more than 30 years, Li Zhengjun consistently worked on the front lines in enterprise management, equipment technology, and IT management. He previously served as Deputy General Manager at Rizhao Port Ore Terminal Company, Deputy Director of Equipment Department at Rizhao Port Group, and Party Committee Deputy Secretary at Lanshan Company. He currently holds the position of Chief Expert at Shandong Port Technology Group and is a Senior Engineer.

Throughout his career, he has been troubled by the ultimate challenge in safety management: the difficulty of enforcing safety protocols. Yet he never stopped thinking about solutions.

At the port terminal, containers stand tall

"Many production accidents stem from improper frontline operations," he deeply understands. "Is there a way to truly enforce safety regulations?"

In autumn 2019, inspiration struck. While observing maintenance work on conveyor belts—a zone stretching dozens of kilometers with sixty to seventy intersecting lines, scattered inspection points, and frequent overlapping tasks—he realized the biggest risks came from workers failing to strictly follow procedures such as power disconnection, lockout-tagout, and confirmation checks.

"If I could see from my office exactly what frontline staff are doing at each step, ensuring critical actions are completed, then safety control would feel solid," said Li Zhengjun.

Someone suggested using WeChat groups to share and display information.

"No," Li shook his head. "Messages get lost. You can’t link steps of a complete operation together. No continuity, no reviewability."

What he wanted was a clear, closed-loop, traceable visual digital "evidence chain."

So he gathered a few young colleagues: "I want every step of a high-risk operation recorded with photos—like candied haws on a stick."

The young team replied: "You can do this with DingTalk."

In October 2019, an initial workflow—the conveyor belt repair process—was born: three key actions before, during, and after operations, each confirmed with photos—"in three, out three." Simple, yet capturing the core of accident prevention.

Initially, some found it cumbersome, but change came fast. Team leaders appreciated seeing real-time conditions, captains valued verifiable records, and employees felt safer. "Before, we always feared misoperation. Now we have photographic proof."

Once the system ran smoothly, Li optimized the process, expanding from conveyor belt inspections to high-altitude work, confined space operations, electrical repairs—and every high-risk procedure generated its own independent "operation record chain." Who entered the site, when locks were applied, whether voltage checks occurred, when personnel evacuated—all became traceable.

Li named this system the "DingTalk Photo Documentation Method." Built upon DingTalk’s robust platform, the name also symbolizes a commitment to making rules “set in stone,” forming a strong oversight system alongside on-site supervision and port video monitoring. Backed by administrative momentum, the "DingTalk Photo Documentation Method" achieved full coverage at Lanshan Company of Shandong Rizhao Port, prompting other units to follow suit. Over ten subsidiaries under Rizhao Port soon adopted the process.

Later, the method received high praise from a renowned safety expert at Peking University and was included in the expert’s publication: "The photo documentation method solves three key aspects of rule enforcement: responsible parties, execution details, and time nodes."

By chance, Li transferred to Shandong Port Group (Qingdao). The group’s safety department noticed his approach: "Can we roll this out across the entire group?" After a three-month pilot, opposition turned into approval. The process was simple, low-cost, and provided clear evidence chains for compliance.

"Safety relies less on words and more on sight—on what can be seen," frontline team leaders told him.

By October 2025, after six years, the "DingTalk Photo Documentation Method" had been fully implemented across 120 companies within Shandong Port Group, covering core port operations and eight non-port sectors. Monthly usage reached 150,000 instances involving 30,000 employees. External companies also began adopting it, with over 2,000 process templates developed.

Driven by safety needs, the first paying customer emerged in 2023. In 2024, another major client from high-end manufacturing joined—CITIC Group’s subsidiary steel company fully adopted the project. The method expanded beyond Shandong, reaching ports in Anhui and Zhejiang, with inquiries coming from Southeast Asia and South America. The system was upgraded and renamed the "Zhi Sun Risk Operation Digital Management Platform," achieving an annual output value of tens of millions of yuan.

Li Zhengjun understands that DingTalk’s success in SOEs stems not just from being a tool, but from enabling an "organizational co-creation mechanism." In the second half of 2025, Tongyi Qianwen was integrated into port management, assisting in image review, report generation, and hazard identification, further boosting efficiency.

When asked about value and impact, Li smiled: "The value of a safety tool can’t be measured by accident statistics. Its significance lies not in 'improvement,' but in 'non-occurrence.'"

He added: "In the past six years, every company using this system has maintained zero accidents and zero injuries."

"We are the gatekeepers of DingTalk"

Baming is Director of Customer Service at DingTalk Service Center. Their mission is to ensure DingTalk truly lands—becoming usable, durable, and trustworthy on countless enterprise screens.

"We are the gatekeepers of DingTalk," he said solemnly.

DingTalk’s growth story is also one of shifting from product-driven to service-driven development. Initially, its clean interface and seamless communication experience quickly won over small and medium-sized enterprises. However, as user numbers grew from hundreds of thousands to hundreds of millions, product complexity and industry diversity became apparent. Even the best features risk being shelved without proper service to explain and implement them. Thus, consensus formed internally: service should no longer be just a customer hotline, but the second leg supporting the DingTalk ecosystem.

November 6, Baming speaking on stage at a Dingsummit hosted by DingTalk in Wuxi

With the arrival of the AI wave, Baming clearly sensed a shift in enterprise sentiment—both anxious and excited. Managers in traditional industries fear technological obsolescence yet eagerly hope to use AI to boost efficiency, seeking ways to optimize warehousing, eliminate middlemen, increase repurchase rates, and improve listing efficiency.

Facing these new demands, the role of the service center evolved. They’re no longer just problem solvers—they’ve become translators of AI implementation, converting complex technologies into solutions businesses can understand and apply. Baming says their job is turning AI into practical tools in enterprises’ hands, not abstract slogans.

There’s a long-standing tradition within DingTalk: managers and developers must personally serve customers. This is part of DingTalk’s co-creation mechanism. Those who write code must hear firsthand users’ confusions and complaints, experiencing real-world usage scenarios. From why a new user can’t find a feature entry, to veteran users questioning new pricing models—these seemingly trivial issues often spark product improvements. This feedback loop keeps DingTalk’s products firmly grounded.

The name of DingTalk’s 8.0 version—“Jue” (Fern)—faced almost no debate during internal brainstorming. As one of Earth’s earliest higher plants, ferns have deep roots, resilient stems, and lush green leaves—symbolizing digging deep into industries, restructuring internal frameworks, and flourishing outward into ecosystems.

October 30, event site of Yiwu–Ningxia Chamber of Commerce in Zhejiang, attendees discussing DingTalk’s AI features

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!