After Alibaba and Tencent's fierce competition for office software entry points, how can other office software companies survive without becoming mere plugins?

If AI-powered office agents become the sole gateway in the future, which companies' value propositions will be undermined?

Last week, Alibaba consolidated its AI office agents. After examining their specific functionalities, I was left with a sense of foreboding. It's evident that within the next year, the landscape of desktop office software will undergo a complete transformation. Browsers, office processing tools, image and video editing software, and even apps like Tianyancha are at risk of losing their entry-point advantage and becoming mere plugins.

Take WPS as an example. Previously, a substantial portion of Kingsoft Office's valuation was tied to WPS, as were its aspirations in the AI domain. Its market capitalization once soared on the back of AI hype, all thanks to WPS. However, if WPS is reduced to a tool summoned by other AI agents in the future, how much value will it retain?

Currently, we can discern three distinct trends.

Firstly, the tech giants will not cede this market.

Secondly, this is destined to be a battle for entry points.

Thirdly, the shift towards skill-based approaches is inevitable.

Why are we confident that the tech giants won't back down? Because this represents the most tangible avenue for cash flow recovery.

From Silicon Valley to Hangzhou, AI large model companies on both fronts are continuously investing in AI office agents. Claude and Codex have emerged as the most sought-after auxiliary tools for professionals willing to pay. In just half a year, Alibaba has evolved its business in this space multiple times, transitioning from the tripartite balance of QoderWork, Wukong, and MuleRun to the current integration into Qianwen Office. Tencent has also consistently invested in Workbuddy.

Every financial report from AI giants in the United States now garners worldwide attention. Can this truly create value?

The willingness to pay for office AI far surpasses that of individual consumers. With companies investing hundreds of billions of dollars annually, there's a pressing need for AI to swiftly generate value and deliver a return on investment. Tencent's progress in AI for individual users has been sluggish, but the evolution of Workbuddy has been remarkable. Tencent has always been attuned to profitability, and it clearly senses the financial potential.

Once the tech giants enter the fray, companies like Kingsoft Office, which previously relied on time-accumulated first-mover advantages, will see these advantages rapidly eroded by substantial capital infusions. Previously, this was a niche market overlooked by the giants. However, once they recognize the potential for traffic entry points, the market dynamics shift immediately.

This brings us to the second point: this is destined to be a battle for entry points.

Imagine a scenario where all white-collar workers, upon entering the office and turning on their computers, first launch their AI office agents and complete all tasks through a dialog box. What is this agent? It would be a super office traffic entry point, sitting atop the operating system, with a value roughly equivalent to the App Store during the mobile internet era.

This entry point could monopolize the attention and behaviors of workers. All software currently opened separately would quickly become plugin-based and unobtrusive.

Traffic is the lifeblood of the online world. Whoever controls the traffic entry points holds the upper hand. In the AI era, this logic remains unchanged. It's akin to offline commerce, where having a storefront on the busiest street guarantees success.

What does it take to become a traffic entry point? Certainly, a product is essential. But when products are comparable, it boils down to who has the deeper pockets. I don't believe companies like Kingsoft Office can compete with Alibaba and Tencent.

It's evident that not only in China but also in the United States, a significant number of software applications will become plugin-based in the AI era.

Why is this plugin-based approach inevitable? That brings us to the final point: skills will become the calling card of the new era.

Human nature leans towards laziness over diligence. The principle of "taking what is readily available" will likely dominate the AI era. The skills crafted by exceptional individuals may represent solutions that many others could never devise in their lifetimes. So why not simply leverage these skills?

As a result, many tasks will become mindless. One can simply summon a skill and then unobtrusively call upon plugins to get the job done, making minor adjustments to details before delivery. This will likely be the norm for office tasks in the future.

In this context, aside from the natural language dialog box of the entry point, most other behaviors will become unobtrusive. Over time, many people may not even be aware of the existence of WPS. This is the predicament.

Finally, let me share my conclusion.

Personally, I'm quite pessimistic about the trajectory of domestic office software in the AI era. Nasdaq has already rendered its verdict on software stocks recently, expressing pessimism and witnessing a significant stock price decline. This logic holds true here as well.

Office, backed by Microsoft, may still have a glimmer of hope. However, I'm absolutely pessimistic about WPS. The reason is straightforward: thorough reform would mean sacrificing cash flow, while not reforming would inevitably result in becoming a plugin.

WPS clearly dares not transform itself directly into a dialog box because its existing paying users come for document processing. This makes it hesitant to commit to the battle for entry points and unable to change user habits. It's similar to how Baidu, when developing Wenxin Yiyan, dared not completely overhaul its existing search box. This is human nature—wanting to have the best of both worlds.

Without thorough reform, compromise is inevitable, leading to a situation where neither end is executed well. It's like boiling a frog in warm water, eventually becoming a plugin for the AI office agents of the tech giants.

In my opinion, rather than this, it would be wiser to transform into an agent oneself, to become one's own plugin. Otherwise, with a current market capitalization of 120 billion yuan, what will sustain it in the future? An AI plugin can't possibly command such a high price, can it?

This is not just an issue for Kingsoft Office but something that all software companies earning money from the B2B sector should ponder.