A 75,000-word exposé has once again thrust DingTalk into the limelight, ultimately sparking a leadership overhaul.

On June 11, Alibaba announced a management shakeup at DingTalk: Chen Hang stepped down as CEO, and Chen Yusen, a tech-savvy manager born in 1992, took the helm.

The timing of this move invites external speculation.

Shortly before, a 75,000-word essay by a former DingTalk employee, titled “Inside DingTalk,” ignited intense internal and external debate. Subsequently, Ma Ruila, former Vice President of DingTalk, published “Outside DingTalk,” confirming his departure and escalating the controversy to organizational management levels.

If “Inside DingTalk” laid bare the wounds of an AI project, “Outside DingTalk” added annotations on management practices.

This leadership transition signals that the issue transcends mere “employee essay” publicity—it’s an organizational and strategic crisis Alibaba must swiftly address.

From a broader perspective, this shift resembles a recalibration of DingTalk amid a pivotal phase for enterprise intelligence transformation.

DingTalk’s unique strength lies in its robust foundation: extensive enterprise coverage, a mature ecosystem of approval, meeting, document, low-code, and industry applications, experience in complex sectors like manufacturing, government, and education, and the technological backbone provided by Alibaba Cloud and the QianWen model.

These assets position DingTalk not as easily replaceable office software but as Alibaba’s closest gateway to real enterprise workflows.

Yet, the stronger the foundation, the harder it is to pivot.

“Inside DingTalk” not only stings DingTalk but also an important piece of Alibaba’s AI-to-B strategy.

Within days, the long essay, resignations, internal responses, and management adjustments unfolded in rapid succession. The AI narrative DingTalk aimed to tell was suddenly overshadowed by another storyline: Why does a platform with the mission of “making work simpler” leave its own employees drained by the pace? Why does a product designed to boost organizational efficiency first find itself mired in controversies over its own organizational efficiency? Why does a project with high hopes from Alibaba expose cracks in its management style when creativity is most needed?

Old strengths are morphing into new pressures. This leadership change marks not the end but the beginning of an inevitable transformation for DingTalk.

Beyond the Long Essay:

The Cracks in AI Transformation Are Visible

The focal point of “Inside DingTalk” is DingTalk’s AI project, ONE.

ONE is not a mere feature upgrade but an attempt to rewrite DingTalk’s entry logic. Previously, users navigated between modules like messages, approvals, meetings, documents, schedules, and to-dos. ONE aims to reorganize these scattered elements, enabling the system to first grasp the user’s work context and then present it more centrally.

This means DingTalk aspires to evolve beyond a tool collection to become a dispatch layer in enterprise workflows, further participating in internal information distribution, task prioritization, and work execution.

In other words, DingTalk seeks to transition from an “office platform” to an “AI-era work entry point.”

Success would redefine DingTalk’s value. However, the office scenario is far more complex than consumer internet.

Consumer products leverage algorithms to boost engagement and recommendation systems to sustain interest. Office products, however, prioritize “accuracy, restraint, and controllability” over “attractiveness.”

In enterprise settings, every information ranking, reminder adjustment, and task aggregation involves responsibility boundaries, permission systems, and collaboration costs. An AI misjudgment isn’t just a UX issue—it could disrupt business progress.

This is ONE’s toughest challenge. It’s not about tweaking a page or button but altering users’ ingrained work habits and existing collaboration norms.

Over the past year, DingTalk has consistently reinforced its AI, agentization, and intelligence strategy externally. From AI spreadsheets and AI note-taking to ONE and Wukong, DingTalk aims to prove it’s no longer just a traditional collaborative office platform but an enterprise work entry point in the AI era.

While ONE’s setbacks shouldn’t be reduced to management issues, the controversy has laid bare the first crack in DingTalk’s AI transformation: the old entry point is aging, the new one isn’t yet established; old strengths remain significant, but new visions remain unfulfilled.

This is DingTalk’s true anxiety.

The previous growth surge in collaborative office tools stemmed from enterprise digitization and remote work popularity. The next wave hinges on whether AI can genuinely transform enterprise productivity.

Once the opportunity window closes, user mindset and enterprise budgets will shift. Whoever first becomes the default entry point for enterprise AI workflows will accumulate data, processes, plugins, agents, and developer ecosystems. Lagging behind risks becoming a backend system summoned by AI tools.

DingTalk can’t afford to stand still, and competitors won’t wait for it to resolve internal issues before forging new paths.

Inside DingTalk:

Management, Product, and Commercialization Challenges

More pointed than ONE’s issues is the organizational state revealed in the long essay.

This isn’t just about an AI product’s success but whether the organization persists with highly certain methods amid exploration filled with uncertainty.

Strategic projects inherently carry pressure, and AI projects aren’t mere feature iterations on mature business lines. They demand trial and error, validation, refutation, and reiteration. They require genuine frontline feedback and allow for clashing judgments within the organization.

If all pressure converges on “proving the strategy correct,” the product team shifts from “seeking user value” to “fulfilling organizational expectations.”

This is why the DingTalk incident escalated from a product critique to a management style controversy.

It resonated because it touches on common challenges large companies face during intelligent transformation.

The faster technology evolves, the more anxious organizations become. The grander strategic slogans, the more uncertainty falls on the grassroots. The more creativity products require, the more management reverts to strong execution, frequent reporting, and oppressive pressure.

For DingTalk, this contrast is stark.

DingTalk’s former slogan was “making work simpler,” but this time, the outside world sees its internal work constantly pressured by pace, reporting, and organization. When a platform designed to enhance enterprise efficiency is questioned about its own organizational health, it backfires on its brand narrative.

Ma Ruila’s “Outside DingTalk” amplifies this contrast. As not just an ordinary employee, his resignation isn’t easily dismissed as personal sentiment but reflects DingTalk’s management state.

The real question he raises is: When chasing the AI era ticket at extreme cost, are employees creating products or being consumed by shifting goals?

This context also explains Alibaba’s Partnership Committee’s rare statement.

Subsequently, the internal response from Alibaba’s Partnership Committee elevated this discussion to group culture levels. The ensuing leadership change signaled to the outside world that Alibaba doesn’t intend to treat this incident as a mere publicity storm.

Especially after Alibaba’s organizational adjustments, business splits, and strategic refocusing in recent years, the group must reanswer: What management style suits the AI era?

DingTalk is eager to change, but urgency breeds the risk of pushing new products with old organizational methods. The more old methods persist, the more creativity suffers, turning AI innovation into slogan-driven feature stacking.

For DingTalk, the leadership change is both a correction and a pressure.

Under the Leadership Change:

DingTalk’s Transformation Enters Recalibration

Equating ONE’s setbacks solely with management issues underestimates the inherent difficulty of enterprise-level product innovation.

With Chen Yusen at the helm, DingTalk’s transformation enters a new phase.

Compared to Chen Hang’s strong product-centric label, Chen Yusen’s background leans toward technical entrepreneurship. He founded cybersecurity company Changting Tech and later led internal entrepreneurship at Alibaba Cloud, developing the AI agent product MuleRun.

Alibaba’s choice of such a young, tech-savvy manager to lead DingTalk suggests, to some extent, that DingTalk’s next phase may focus not just on product reconstruction but on technology, agents, and enterprise-level execution capabilities.

This aligns with Alibaba’s broader strategy.

Alibaba’s current focus is clear: one end is consumer spending, the other is AI + cloud. The former drives cash flow and user scale, while the latter fuels future growth expectations. Alibaba Cloud, QianWen, BaiLian platform, enterprise-level agents, and industry solutions collectively form the foundation of Alibaba’s AI-to-B strategy.

However, for large models and cloud services to truly penetrate enterprises, they can’t stop at APIs and computing power. Without integration into real enterprise business systems, translating technological narratives into commercial results is difficult.

This is where DingTalk’s role lies.

It has enterprise organizational relationships, employee identity systems, approval flows, meetings, documents, low-code platforms, and application ecosystems. It’s close enough to real enterprise work scenarios and Alibaba Cloud’s technological foundation. Theoretically, DingTalk is the key intermediary layer for Alibaba to translate model capabilities into enterprise productivity.

But this also means DingTalk can no longer be just a collaborative office product.

It must help Alibaba Cloud find more enterprise-level AI implementation scenarios, inherit QianWen’s model capabilities, drive agent ecosystem distribution, and prove to enterprise customers that AI investments yield real returns.

DingTalk has become a crucial validation ground for whether Alibaba’s AI-to-B commercial closed loop can succeed.

In this context, ONE resembles the first-layer experience for users entering DingTalk, while Wukong represents what future enterprises are willing to entrust to AI for completion. The former is more of an entry point transformation, while the latter is closer to an execution platform.

By embedding into organizations and businesses, DingTalk could potentially upgrade from a collaborative office platform to an enterprise agent distribution and execution platform. This would directly help Alibaba Cloud increase enterprise customer stickiness and drive model invocations, cloud resource consumption, and industry solution sales.

However, commercializing enterprise-level AI is no easy feat.

Enterprise customers care not just about feature advancement but also data security, permission isolation, system stability, audit compliance, implementation costs, and ROI. Whether an agent can write summaries matters less than whether it can safely execute tasks in complex organizations, reduce labor costs, and improve process efficiency—these are the key drivers for enterprise payment.

Therefore, DingTalk can’t just tell AI stories; it must swiftly deliver verifiable commercial results. This is also an urgent issue for Alibaba as a whole.

Without transformation, the imagination space of old office collaboration models will shrink. However, transforming too hastily may amplify organizational and product issues. DingTalk can’t afford to be slow or chaotic; it must change but not at the cost of consuming team and user trust.

This is the first dilemma Chen Yusen must confront after taking charge.

After Half-Time:

From Management Efficiency to Productivity Platform

The competition DingTalk faces is no longer just about office software.

WeChat Work’s strength lies in its connection to the WeChat ecosystem. It’s not just an internal office tool but also infrastructure for enterprises to connect with customers, upstream and downstream partners, and external social relationships. For industries like retail, education, healthcare, finance, and lifestyle services, WeChat Work’s core value lies in linking employees, customers, and private domain operations, vying for the entry point to enterprise external relationships.

Feishu’s advantage lies in product experience and knowledge workflows. It entered through scenarios like documents, meetings, multidimensional tables, and project management, making it popular among internet, tech, R&D-oriented, and knowledge-intensive organizations. Feishu attempts to integrate documents, spreadsheets, knowledge bases, business systems, and AI development tools, vying for the entry point to enterprise knowledge and collaboration.

DingTalk’s strength lies in organizational coverage, business processes, and low-code ecosystems. It has stronger penetration in traditional enterprises, mid-to-large organizations, manufacturing, government, and education, making it easier to reach deep into enterprise processes. DingTalk vies for the entry point to enterprise organizations and processes.

The rivalry among the trio primarily revolves around three distinct strategies: WeChat Work leverages relationship networks, Feishu focuses on productivity tools, and DingTalk centers on organizational digitization and business process optimization.

However, the advent of AI is reshaping the competitive landscape.

Traditionally, collaborative office platforms competed based on the comprehensiveness of their instant messaging (IM), meeting, document, approval, spreadsheet, and project management capabilities. Looking ahead, the competition will shift towards who can seamlessly integrate AI into the core work processes of enterprises. WeChat Work can integrate AI into customer service and private domain operations, Feishu can weave AI into knowledge creation and business system development, and DingTalk must incorporate AI into organizational management and process execution.

This presents both an opportunity and a challenge for DingTalk, as its foray into management processes brings it closer to the delicate balance between "efficiency" and "control."

The real conundrum lies in the fact that while past success was rooted in management efficiency, future growth must stem from productivity enhancements.

These two objectives do not always align. Management efficiency prioritizes visibility, controllability, and traceability, whereas productivity improvement emphasizes autonomy, collaboration, creativity, and outcomes.

For DingTalk to sustain its enterprise procurement advantage, it must also alleviate the burden on employees.

This issue is intricately linked to the deeper pressures DingTalk faces in its commercialization efforts.

The Chinese enterprise SaaS market has always been challenging. Many enterprises are willing to use tools for free but are reluctant to pay premium prices for standardized software. They are open to investing in customized systems but may not necessarily commit to long-term subscriptions. Despite DingTalk's vast user base, converting this scale into high-quality revenue requires more AI-driven scenarios that directly generate business value.

In other words, DingTalk cannot merely settle for "enterprises using it"; it must address the question of "why enterprises are willing to continuously pay for advanced capabilities." It needs to transition from "selling tools" to "selling outcomes."

Whether a sales agent can boost lead conversion rates, a customer service agent can reduce labor costs, a financial agent can shorten reimbursement and reconciliation cycles, an HR agent can enhance recruitment and training efficiency, or a manufacturing agent can minimize losses in workshops, supply chains, and project management—these are the tangible reasons enterprises are truly willing to invest.

Following the leadership transition, DingTalk still has opportunities.

It boasts a robust organizational foundation, the support of Alibaba Cloud and QianWen, a thriving low-code and enterprise application ecosystem, and extensive experience in enterprise-level scenarios.

DingTalk needs to clarify:

What genuine problems does AI solve?

Merely adding intelligent summaries, searches, and Q&A features to all functionalities will not suffice to break through commercialization barriers. Enterprise customers will not indefinitely pay for concepts; they seek cost reductions, efficiency improvements, quality enhancements, and growth.

Simultaneously, can the organization sustain long-term innovation?

DingTalk's future hinges not on the quantity of AI products it launches but on its ability to reaffirm that a platform designed to enhance enterprise efficiency must first possess the capability to maintain its own organizational creativity.

With Chen Hang stepping down and Chen Yusen taking the helm, DingTalk stands at a new crossroads.

This transformation will not be halted by a mere essay or automatically achieved through a leadership change. DingTalk must navigate a challenging identity shift within the AI opportunity window.

Of course, the true crisis is not that the outside world perceives the problems but that no new solutions emerge after these problems have been identified.