This perspective stems from an observation that in the domestic collaborative office space, there is an overemphasis on AI, yet in practical applications, the main functions of tools like Feishu and DingTalk (such as online documents, multidimensional spreadsheets, and check-in systems) actually have little to do with AI.
In the past two years, there has been a surge of new entrants in the software field, with most startups targeting AI. Whether it's search, generation, or conversation, to name a few, products are generally based on some large model to leverage their application value, and these companies have been given high valuations as a result.
However, in the collaborative office space, current AI applications seem to be more of a gimmick than a real selling point. This is because for most collaborative office scenarios, AI has not brought significant efficiency improvements or cost reductions. Real-world customer companies also indicate that their actual demand for AI is limited, and many so-called AI features have not been widely accepted or recognized.
Imagine, in some cases, introducing AI actually makes the workflow more complicated, such as after using AI for information retrieval, manual verification of its correctness is still required. Since the intelligence of AI depends on the accuracy of the original data, user trust in AI is limited, which restricts the widespread application of AI in collaborative offices.
Advertisement
In fact, we have also conducted some research in the collaborative office space abroad and found that the problems faced domestically and internationally have some commonalities to a certain extent.
Take Zoom, for example. Three or four years ago, it was an undisputed star company. The global pandemic has made significant changes to people's lifestyles and work habits. Zoom, which focuses on online video conferencing, is simple, easy to use, and compact, becoming the preferred tool for remote work, learning, and socializing, with its market value once approaching $160 billion.
Zoom's success has shown the market the imagination of collaborative office spaces, and peers like Microsoft Teams, Slack, and DingTalk, both domestic and international, have also seen an uptick during the same period.
However, after the highlight period, Zoom's development curve is not as attractive:
As of now, the stock price and market value are significantly different from the past. There have been speculative comments, "A few days ago, Zoom showed a rebound trend, and I thought it might make a comeback. However, after a few days of observation, it did not truly change the previous downward trend."
Why is Zoom showing a downward trend? The reasons can be attributed to the fading of the pandemic dividend, increased market competition, slowing customer growth, and increased marketing and R&D expenses, among other factors. This also gradually makes people realize that what determines a company's growth curve is not only how the product is but also its commercialization path.It's interesting to note that Yuan Zheng, the founder of Zoom, does not view the explosive growth during the pandemic as a positive development. To cope with the surge in users, Zoom had to rapidly expand its workforce, which impacted the company's culture. Competitors also accelerated their support for remote work and video conferencing during the pandemic, increasing competitive pressure on Zoom.
"Without the pandemic, Zoom might have developed better," Yuan Zheng believes that the current situation for the company does not mean it has returned to a reasonable range. With a small and beautiful product but a low moat, Yuan Zheng pins his hopes on AI, which aligns with domestic players' goals.
According to media reports, the product, originally focused on video conferencing, now advocates against using video meetings. AI has been applied to almost all product lines, with the ultimate goal of promoting the use of AI avatars to handle most work, attend meetings, make decisions, and even reply to emails. Yuan Zheng also plans to develop Zoom into a comprehensive platform, launching a collaborative platform called Workplace, as well as tools for documents, email, and more.
In his view, AI has played a role in reshaping Zoom's moat, with keywords being: completely free, not using user data to train AI, and low cost. Their AI strategy not only focuses on current performance and cost but also maintains competitiveness in the ever-changing AI field by combining multiple technology sources, using different third-party models for various scenarios and swapping them, as well as the reasoning process.
Zoom has introduced several generative AI features, including Zoom AI Companion and Zoom Docs, which are embedded in various Zoom tools such as Zoom Meetings, Team Chat, and Whiteboard. These features aim to enhance enterprise productivity and inter-team collaboration efficiency.
From these examples, most AI features serve as an auxiliary role to existing functions, but "free" has helped Zoom build a deeper moat.
On one hand, in the fiercely competitive SaaS market, providing additional free features can increase user stickiness. On the other hand, offering AI features for free allows users to try and use these new features without barriers, attracting more users in the long term, and then achieving profitability through other means (such as value-added services, enterprise subscriptions, etc.).
Take Zoom's Contact Center as an example; it is a modern customer service solution, and the biggest difference from other products is its combination of video calls and AI technology. The decision-makers who purchase Contact Center are different from those who previously bought video conferencing products, but since they are also Zoom users, their trust in Zoom helps with Contact Center sales.
Another example is the star software Notion, which has recently undergone significant updates, especially in the comprehensive restructuring of its AI features. The upgraded sections mainly focus on three aspects: improved interaction of conversational robots, optimization of writing assistance functions, and revitalizing the knowledge base, making AI assistants interact more naturally with users.
The newly upgraded Notion AI can access the knowledge base across different pages and applications. This cross-application access capability allows users to more conveniently obtain and utilize various resources, thereby enhancing overall productivity.As one of the earlier products to integrate with ChatGPT, Notion AI was fully opened to all users at the beginning of this year and officially made AI capabilities a separate paid product. Its pricing strategy is $10 per person per month, or $8 per month if paid annually. Notion AI can assist users with a variety of tasks such as writing, editing, summarizing, and more. It also offers cross-platform search functionality, seamlessly connecting with applications like Slack and Google Drive to quickly aggregate information from multiple sources.
Reports indicate that as of 2023, Notion had reached 30 million users, with approximately 4 million paying customers. The initial internal assessment was that Notion AI would bring in an additional $10 million to $30 million in Annual Recurring Revenue (ARR) within a year. However, this target was achieved in just one month, implying that Notion AI might ultimately generate $100 million in ARR annually. Moreover, Notion AI quickly attracted a large number of users after its launch, surpassing 4 million in a short period of time.
It has to be said that this level of commercial growth is beyond the reach of most domestic office software.
A typical example is DingTalk, despite having a massive user base (reaching 700 million users and 25 million enterprise organizations by the end of 2023), the actual number of paying users is quite small. Previously, CEO Ye Jun stated that less than 1% of enterprises using DingTalk might actually pay.
To promote commercialization, DingTalk has integrated Alibaba's "Generalized Q&A" large model, using AI to redo the product, completing the AI-ization of multiple product lines and features.
For instance, by simply pressing the "/" key, it intelligently generates most of the work content in the workplace, including document creation and chat summary generation. In addition to offering a free standard version, DingTalk has also introduced annual fee service options such as a professional version, exclusive version, and proprietary version. The introduction of AI features by DingTalk is primarily aimed at adding value to these paid versions, making users feel that advanced features are worth the investment.
Furthermore, DingTalk also increases revenue through platform commission sharing and hardware open License authorization models. By leveraging AI technology to provide personalized services for enterprises, such as intelligent analysis of enterprise data and custom workflow, DingTalk believes that enterprises will be more willing to pay, and the relationship will be closer.
However, insiders have learned that the actual number of customers willing to pay for this is not yet large, and DingTalk has not yet achieved profitability.
On the other hand, Lark, which is also integrating AI into multiple products such as multi-dimensional tables, field shortcuts, and dashboards to enhance the intelligence level of the products, has launched its own AI assistant "My AI" (Lark Intelligent Partner). Users can ask it to write work summaries, meeting records, and activate it for collaborative Q&A during office hours.
In terms of specific commercialization, with the continuous supplementation of AI features, there may be value-added services or subscription models based on AI features. For example, providing users with a higher-level AI assistant, Lark is more inclined to attract customers by offering professional services and customized solutions.Overall, foreign collaborative office tools often adopt a pay-per-feature model, allowing users to select the features they need and pay accordingly. In China, despite tools like DingTalk and Lark actively exploring commercialization paths, they still face numerous challenges in commercialization. This is due to differences in the market environment between domestic and international markets, as well as varying perceptions of the value of collaborative office tools and willingness to pay among users.
According to existing reports, in the international market, user privacy issues have had a significant impact on Zoom's AI deployment. However, as Zoom has taken a series of measures to ensure the security and privacy of user data, this issue has been somewhat alleviated.
Has Zoom achieved greater growth due to AI? Based on data from multiple sources, we have found that the integration of AI technology has indeed increased its enterprise sales. For example, in the third quarter of the fiscal year 2024, Zoom's total revenue increased by 3.2% year-on-year, with enterprise revenue growing by 7.5% year-on-year to $661 million. After integrating AI technology, Zoom's quarterly customer churn rate has narrowed significantly, showing a higher customer retention rate.
Why is there such a big difference in commercialization between Zoom, Notion, Lark, and DingTalk in the domestic and international collaborative office fields, despite the same AI deployment? Is it because the AI capabilities of domestic software are not strong enough, or is it because the market's expectations for AI are too high?
This is just one aspect. Upon comparison, AI is indeed a powerful technology, but its role in collaborative office settings is not omnipotent, and its current use cases are limited, mainly appearing in the form of intelligent assistants. As Yuan Zheng said, a good product does not necessarily correspond to good commercial value; R&D costs, business models, and payment environments also play a decisive role.
Many collaborative office companies have invested a lot in R&D for AI, but they have not achieved a considerable ROI. Instead, they have become project providers offering digital solutions to enterprise clients or software outsourcing companies.
Returning to the essence of business, customer needs and market positioning remain key factors in determining product success. Although AI technology has developed rapidly, bringing unprecedented changes to the industry, the core principles of business have not changed. However, excessive investment in AI can sometimes indeed cause companies to lose their way and wonder "what they should do."