How can established products rediscover their Product-Market Fit (PMF) in the AI era? This is a pivotal question for DingTalk, Alibaba's leading business collaboration platform.

Financial indicators suggest that DingTalk's commercialization efforts have gained momentum this year. In 2024, the platform announced an Annual Recurring Revenue (ARR) of $200 million for the first half of fiscal year 2025 (April to September 2024), with a target of achieving breakeven in 2025. However, Alibaba's financial reports for the past two quarters (October 2024 to March 2025) omitted any mention of DingTalk's profitability.

As a flagship product in Alibaba's AI portfolio, the absence of profitability data from financial reports hints at challenges in DingTalk's path to breakeven. Despite its apparent alignment with AI trends, particularly the surge in popularity of AI agents in China, the recent leadership change at the end of the first quarter will inevitably impact DingTalk's commercialization trajectory.

DingTalk's current situation mirrors the market's reaction to Alibaba's financial reports – solid financial figures that fall short of lofty market expectations.

With the return of No Recruit, DingTalk's founder, at the end of the first quarter, the ongoing strategy under Ye Jun will likely shift. Consequently, the breakeven target set for 2025 by Ye Jun may be delayed.

Will No Recruit's bold approach help DingTalk find its PMF in the AI era? This raises questions about balancing individual creativity with organizational rules and workflow dynamics in the AI context.

01. Will the Breakeven Point Be Delayed After No Recruit's Return?

No Recruit's return as DingTalk's new leader marks a decisive start.

Social media reports indicate that DingTalk has tightened attendance management, introduced morning meetings, evening summaries, and strict working hours. Even the Python exam for R&D teams, notified by DingTalk's CTO Yili, is being conducted. Beyond these institutional changes, leaked information suggests that No Recruit's vision for DingTalk is more radical than Ye Jun's.

Previous analyses, such as "No Recruit Returns to DingTalk, Alibaba's Correction to 'Strategic Correction'" by Xinlichang, highlighted the stylistic differences between No Recruit and Ye Jun: No Recruit's "productism" versus Ye Jun's "operational thinking," representing distinct paths for B2B businesses – the former emphasizing tool value, the latter pursuing ecological binding.

No Recruit's actions since his return confirm this distinction. On his first day back, he met with all product research and design staff above the P7 level to assess the product landscape. Each product and design team was tasked with a comprehensive review of the product experience chain.

The product review focused on two areas: scenario sorting, where DingTalk identified the N minimum Minimum Viable Product (MVP) scenarios for each product function, and paywall analysis, examining the number of payment checkpoints from entry to deep integration.

Reportedly, several product features were removed, and multiple paywalls were dismantled to enhance product and customer experience. This indicates that No Recruit prioritizes these aspects upon his return.

Before this, DingTalk was still navigating the aftermath of the integrated cloud and DingTalk strategy, adhering to Ye Jun's concept of using Platform as a Service (PaaS) as the commercialization foundation.

DingTalk termed this concept the PLG (Product-Led Growth and PaaS-driven) + SLG (Sales-Led Growth and Service-driven) model. Until November 2024, Ye Jun reiterated at the DingTalk Autumn Summit that "DingTalk has found a commercialization path suitable for China's B2B industry," referring to the PLG+SLG path.

At that time, DingTalk disclosed an ARR of $200 million for the first half of fiscal year 2025 and predicted breakeven in 2025. Notably, ARR reflects annual customer contracting, and given the high barriers of the PaaS market and a strong sales strategy, renewal and new contracts are inherently predictable. Once a business model is established and a significant ARR scale is achieved, profitability can be better estimated.

In other words, Ye Jun's 2025 breakeven estimate was deeply tied to the PLG+SLG strategy. This strategy may now undergo vertical extension, as discussed below.

However, it's undeniable that enhancing customer experience by removing features and paywalls could delay DingTalk's breakeven point.

From a group perspective, DingTalk's contribution to Alibaba's revenue remains insignificant. The $200 million ARR in half a year is minuscule compared to the $14.1 billion revenue from DingTalk's "All Other" segment in the same period.

From a revenue structure standpoint, the PLG+SLG model tends to focus on Key Accounts (KAs) with high customer unit prices and customization needs.

Previously, DingTalk disclosed its paid customer structure. As of March 2023, DingTalk had 100,000 paid enterprises, with 58% being small and micro-enterprises, 30% medium-sized enterprises, and 12% large enterprises. Although more enterprises are adopting cloud and digital management, small and medium-sized enterprises are gradually growing into large enterprises.

In summary, while the breakeven point may be delayed, DingTalk must now find a new PMF in the AI era to potentially tap into a larger market in the future.

02. From SaaS, PaaS to XaaS: PMF in the AI Era Hinges on 'Returning to Entrepreneurship'

When No Recruit left DingTalk in 2020, after more than six years of establishment, the platform lacked innovations in industry technology, products, or commercialization concepts that could significantly boost group revenue. Product operation was thus DingTalk's top priority.

However, with the advent of large AI models in late 2022, DingTalk, which could not support group performance, needed to find a new PMF in the AI era to tap into a larger market. This aligns with No Recruit's philosophy.

It took Alibaba and DingTalk another two years to recognize this. During this period, the prevalence of large models and cloud services gave DingTalk hope in adhering to the PaaS foundation. However, the rapid popularity of AI platforms like DeepSeek and Agent earlier this year accelerated AI commercialization, necessitating a reconsideration of underlying strategic logic.

Since his return, No Recruit has not completely discarded the value of the PaaS foundation strategy but recognizes that connecting DingTalk's PMF in the AI era requires more than just PaaS.

Previously, DingTalk had already demonstrated the idea of providing integrated digital services through XaaS: initially defined as Software as a Service (SaaS) and later viewed PaaS as the service foundation after integrating with Alibaba Cloud. With the advent of AI PaaS in the first year of the AI large model era, and the decoupling of cloud and DingTalk, along with exploring Agent on the DingTalk large model, DingTalk's business model has evolved into an XaaS style – a fusion of SaaS, PaaS, and MaaS (Machine Learning as a Service).

From an enterprise digitization perspective, offering mixed XaaS services was already a trend. The essence is that standalone PaaS, SaaS, or MaaS models can no longer meet most enterprises' needs for efficient digital transformation, and the popularization of AI technology has further accelerated this trend.

Under the trend of integrated XaaS services, the first step is to identify common enterprise needs. No Recruit's approach to uncovering these needs is through "co-creation."

In the first week after his return, No Recruit led his team on an intensive visit and survey to Beijing, Guangdong, and East China, engaging in "co-creation" with DingTalk customers.

Media reports suggest that during customer visits, No Recruit preferred a more "undercover" approach to hear firsthand voices: he did not bring any frontline service teams and did not fully refer to the customer survey list provided by the business department.

This near "re-entrepreneurial" state aligns with DingTalk's initial entrepreneurial phase.

The confidence behind this return to "co-creation" stems from three fundamental payment, product, and AI logics:

1. The core discrepancy in B2B business models and product iteration: the payment decision-maker is often not the first product user.

2. The voices of first B2B product users are often difficult to convey upwards (not due to fear but more often a mindset of "the company is not mine" and adherence to rules).

3. AI product iteration most needs the authentic voices of first users. Conversely, those needing AIGC or Agent the most are individual users with the potential to become super individuals.

In fact, these three logics explain why smaller platforms like Feishu can find PMF faster in the AI era. Compared to DingTalk, Feishu's product philosophy aligns closer with individual creativity, while DingTalk's overall product logic aligns more with organizational rules and workflow dynamics. Both concepts exist in both DingTalk and Feishu, as they are crucial in iterating collaboration methods for organizational and enterprise management.

Additionally, WeChat Work's private domain marketing drive is another aspect that can be temporarily ignored. Historically, Feishu and DingTalk's product-level explorations balanced individual creativity with organizational rules and workflows. A case in point is the collective adoption of Big Five AI Models by both platforms last year.

After No Recruit's return, how will he balance DingTalk's product genes driven by organizational rules and workflows with the enterprise needs driven by individual creativity in the AI era?

The answer lies with enterprise managers and frontline users where these two forces can compete on equal footing. This seems to be an issue at the level of enterprise management philosophy iteration.

03. Conclusion

Beyond the aforementioned points, DingTalk's deeper genes stem from its understanding of organizational efficiency within a large group like Alibaba and its profound knowledge of Alibaba's survival strategies in serving enterprises.

Therefore, DingTalk still has the momentum to find a balance between individual creativity and organizational rules and workflow dynamics in its product philosophy, thereby discovering its PMF in the AI era.

As Eric Schmidt, former CEO of Google, conveyed at TED: AI progress is not linear but exponential.

At least for now, DingTalk needs to focus on finding greater acceleration rather than finding it faster.

*The lead image and illustrations in this article are sourced from the internet.