With the popularization of technology, the problem of homogenization of AI search products has gradually emerged. This article will explore how AI search products can stand out through differentiated competition with the help of DeepSeek.
AI technology is “exploding” the search industry.
The valuation of American company Perplexity soared 17 times, and OpenAI and Google have successively upgraded their search functions; Domestic Internet giants such as Baidu, WeChat, and Weibo have also implanted AI search into the main site APP.
According to QuestMobile data, as of March this year, the total number of monthly active users of AI search (the three major terminals) has exceeded 600 million, becoming one of the four major AI application scenarios.
The competition between players is characterized by “lightweight breakthrough”.
In March 2025, 70% of the top 10 applications on the track will adopt the In-App AI model, and rely on the huge traffic of the main application to rapidly popularize intelligent search, Baidu (AI search), Byte (Douyin AI search and Toutiao AI), Tencent (WeChat AI search and Tencent News AI), and Weibo smart search, occupying a dominant position in the industry.
What does a product manager need to do?
In the process of a product from scratch, it is not easy to do a good job in the role of product manager, in addition to the well-known writing requirements, writing requirements, writing requirements, there are many things to do. The product manager is not what you think, but will only ask you for trouble, make a request:
View details >
In terms of monthly active users, Baidu AI Search leads the market, while Weibo Smart Search has become the fastest growing platform with a month-on-month growth rate of 300.2%.
Source: QuestMobile report
The emergence of the open source large model DeepSeek has greatly lowered the technical threshold, and the platform for choosing access stands on the same starting line, and many AI search products have been launched from project approval to launch in just a few months.
However, the “equal power of computing power” brought about by the open source model has also led to the homogenization of products.
In such an environment, it has become more difficult for AI search products to successfully break through, and each company must also make great efforts to differentiate.
What ultimately determines the market direction is always the real experience of users. The market will give the most direct judgment through users’ choices: which AI search products are really easy to use and which can be used by users continuously.
What are the different answers given by each family to ask the same question?
In recent years, AI searches on the market have become increasingly abundant, and while self-researching, many players have chosen to access DeepSeek to enhance their search capabilities.
Connected Insight noticed that the data “fed” to AI by various platforms, even subtle changes, can affect the overall performance of the product.
In order to verify this, we selected four mainstream AI search products for comparison and testing: Baidu AI Search, Douyin AI Search, WeChat AI Search, and Weibo Smart Search, among which Baidu AI Search, WeChat AI Search, and Weibo Smart Search have been connected to DeepSeek. We found that the exact same question will be answered with obvious differences on different platforms.
At present, the core users of AI search can be roughly divided into two categories: one is users who need professional and in-depth content, and the other is ordinary users who solve daily problems.
Depth users are most concerned about the professionalism and reliability of search results.
Taking the question of “how artificial intelligence will change the structure of the job market in the next decade” as an example, the performance of the four platforms has its own characteristics.
Among them, Baidu AI’s quantitative indicators are the most intensive, with a total of 42 data citations, but some lack source explanations. Weibo Intelligent Search and WeChat AI are more authoritative in the selection of sources, citing the research results of MIT, Stanford and other institutions, the views of corporate experts, and authoritative media reports, etc., which are more rigorous.
In contrast, some of the points in Douyin’s AI answers do not indicate the data source, but mention more similar technology landing scenarios in industry applications.
Starting from the needs of knowledge-based users, Weibo Smart Search and WeChat AI are better in terms of data accuracy and content rigor. However, if you want to do more in-depth research, AI search will not replace academic websites like CNKI for the time being.
In order: Baidu AI search, Weibo smart search, WeChat AI search and Douyin AI search
When obtaining more time-sensitive social hot information, Weibo intelligent search performed better among several products.
For example, when querying real-time data such as “gold price trend”, Baidu AI’s answer structure is clear and still cites a large amount of data. Douyin AI excels at capturing real-time data, such as the day’s gold price and short-term forecasts. WeChat distinguishes between international and domestic gold prices and conducts multi-dimensional comparisons.
It is worth noting that Weibo Zhisou’s answer quoted social media dynamics to capture market sentiment, and at the same time, it also referred to the short video content released by authoritative media such as CCTV Finance, and the real-time data was updated in a timely manner.
In order: Baidu AI search, Douyin AI search, WeChat AI search and Weibo smart search
In addition, we also tested these four AI search products from the perspective of ordinary users to see if they could be used to solve the problem of universality in life and entertainment.
Although these questions are not difficult to answer, users also have special needs. It’s like when netizens “eat melons” now, they are no longer just just watching the flowers, but paying more attention to the complete context of hot events and pursuing a true and reliable chain of evidence.
We chose the entertainment topic of “Kei Tanaka cheated on Meiyu Nagano” and observed the different answers of the four platforms according to user needs.
Among the four products, Douyin AI’s answer is relatively brief, in contrast, Weibo Zhisou’s answer is the most detailed, citing a lot of specific data and social media sources, listing the development of the event in detail by timeline, and summarizing the key points of controversy and impact.
In terms of timeline sorting, each product also has differences. Although Baidu AI and Douyin AI both have clear time nodes, Weibo Smart Search and WeChat AI list the events of each key date in more detail. Among them, Weibo Smart Search also made a phased summary, from the revelation to the follow-up development, so that users can fully understand the beginning and end of the incident.
In terms of evidence chain, Weibo Zhisearch mentioned surveillance videos, LINE records, photos and other evidence, and discussed the controversy over its authenticity, while Baidu AI and WeChat AI also mentioned the evidence, but did not analyze the credibility of this evidence in depth. Douyin AI briefly mentioned the contradiction of the evidence, but did not expand it.
In addition, in terms of social impact and follow-up results, Weibo Zhisearch mentioned brand withdrawal endorsements, film and television project changes, legal risks, etc., which are relatively comprehensive. Baidu AI also mentioned a similar impact, while WeChat AI and Douyin AI have relatively little information in this regard.
In order: Weibo Smart Search, Baidu AI Search, Douyin AI Search and WeChat AI Search
In terms of life scenarios, we tested common needs such as “Yunnan tourism planning”. All four platforms can give detailed itinerary suggestions, including attraction arrangements and transportation methods at different times of the day.
In general, the biggest advantage of AI search is that it can answer user needs that are difficult to meet in traditional searches in a simple and clear way.
This is the key to AI search’s rapid rise in just two years and gradually becoming the mainstream way for people to obtain information.
“Partial students” seize the AI search ecological niche
With the gradual popularization of AI search, the traffic distribution pattern of the entire Internet era may be rewritten.
In the era of AI search, users no longer need massive information, but fast and accurate answers, and the “big and complete” of traditional search is no longer an advantage. Although traditional search players are still actively transforming, they have to face the reality of traffic differentiation.
According to MoonFox data, the number of AI search users is expected to increase by 240% year-on-year in 2025, of which 65% of the growth will come from emerging platforms.
This market gap is where the opportunity lies for content platforms.
Almost all content platforms are actively deploying, giving full play to their partial advantages, and striving to set up a flagpole in the next round of traffic redistribution to maintain their own ecological niche.
For example, Weibo is firmly occupying the important ecological niche of hot topic communication through the model of “hot search + smart search”.
As the birthplace of the most influential hot spots in China, Weibo hot search is not only a gathering place for new things, but also an important vane in the Internet public opinion field.
Unlike other platforms that focus more on long-tail content, Weibo can quickly capture and amplify the most time-sensitive news, entertainment hotspots and other topics, and enhance user stickiness through special functions such as celebrity interaction and fan community.
The launch of Smart Search amplifies the existing advantages of its hot search ecosystem: under the traffic guidance of hot searches, users can jump to the smart search interface by jumping to the “Ai Smart Search” tag or clicking on the smart search interpretation card under the entry.
With the help of this coherent experience, hot searches attract a large number of users to click, guiding them to gain an in-depth understanding of the full picture of the event through intelligent search; The conclusions output by intelligent search after integrating information can in turn promote the further fermentation of hot search topics. The combination of the two further strengthens the hot ecological niche of Weibo.
Source: Weibo
In addition, Weibo is promoting cooperation between smart search and mobile phone manufacturers to attract new users.
According to the Weibo earnings call, Weibo has reached cooperation with a number of domestic mobile phone manufacturers to pre-install the hot search list on the mobile phone, and some of the content directly quotes the results of the smart search. In the future, if the search involves news content, Weibo will encourage partners to give priority to citing Weibo smart search results, and guide users to click on the hot search list to jump to the Weibo client.
Similar to Weibo, Xiaohongshu also focuses on its areas of expertise, especially emphasizing the core positioning of “life experience sharing”.
Taking Xiaohongshu “Diandian” as an example, it is positioned as a life search assistant, which precipitates the real life experience of hundreds of millions of users on the platform into a corpus and presents it to users in the form of pictures and texts.
When users search for “commuting outfits”, they can not only obtain multiple matching solutions, but also screen out suitable blogger demonstration content according to their own height and weight conditions.
At present, content communities represented by Weibo and Xiaohongshu are seizing the minds of users. Nowadays, when users mention search, the presence of traditional search websites has been further reduced, and platforms such as Weibo and Xiaohongshu are becoming important search entrances.
It is reported that at the beginning of this year, the effect of Weibo Smart Search was significantly improved after accessing DeepSeek, with monthly active users exceeding 30 million in March, and after the full launch in April, the data is expected to further increase, driving the overall search DAU and query volume.
Although user habits are gradually changing, there is still a lot of room in the AI search market for players to explore.
We also found that other platforms are also building moats in their respective fields.
For example, Zhihu, which focuses on high-quality knowledge sharing, emphasizes that the product has a huge academic knowledge base and is aimed at academics and knowledge workers.
At present, Zhihu Direct has accessed 50 million academic literature, and users can not only summarize the literature views through search, but also read any paper.
The source of the picture knows how to go directly
Bilibili, Douyin and other players mainly emphasize the advantages of short video content. Among them, Bilibili focuses on video content in the two-dimensional culture and knowledge category (mainly PUGC), and will give priority to providing relevant video links when answering user questions. Douyin, on the other hand, focuses on entertainment and trendy content, and when its AI search generates answers, part of the information comes from the intelligent refinement of short video content, and the main site video link will be attached to the end of the answer.
On the whole, the competition of AI search is essentially a competition of content.
Every search request from a user is actually calling the platform’s content database. The richer and higher the quality of the data, the more comprehensive and accurate the information users get. The key to determining the difference in user experience lies not only in the technology itself, but also in the thickness and quality of content resources accumulated by each platform over the years.
AI search has become a battleground for military strategists
As the new “traffic entrance”, AI search is becoming ubiquitous.
From WeChat dialog boxes, Weibo hashtags, to input method interfaces, any entry can become a trigger point for AI search. After this “non-inductive search” is quietly integrated into every corner of life, information acquisition becomes more natural.
As more and more users search through non-traditional portals, the traditional search box pattern is being replaced.
Recently, Apple’s senior vice president Eddie Ku publicly stated that Safari’s search volume declined for the first time in April this year, and original users are more inclined to use AI tools to obtain information.
This trend has led Apple to “actively consider” whether it will end its long-standing relationship with Google. Affected by this news, Google’s parent company Alphabet experienced a stock price plunge on May 7.
This shows that traditional search engines that do not transform quickly enough are likely to be eliminated in the future.
In this transformation, the In-App AI model has shown strong vitality.
Based on its own data advantages, the content platform has developed its own unique AI search services. These “embedded” searches, although they do not exist independently, have a lot of influence, so that the platform’s content ecology has gradually evolved into the form of “content + service”.
In fact, AI search is evolving into an “ecological entrance”, connecting all aspects of the platform ecosystem and connecting multiple scenarios such as content, services, and transactions.
WeChat search has opened up the ecology of official accounts and mini programs; Douyin search integrates short videos and local life; Taobao search connects the shopping scene.
On Xiaohongshu, search results can no longer be just a simple list of notes, but a “content square” that integrates products, discussions and live broadcasts.
Weibo connects hot search and smart search to the platform’s traffic pool, and the two-way traffic flow allows users to change from “looking for information” to “shopping for content”, and the stay time is greatly increased.
The capital market is full of expectations for the prospects of AI search. As the most out-of-the-circle AI search product, Perplexity’s valuation has exceeded 100 billion yuan, and its business models such as subscription system and advertising services are gradually gaining market recognition.
Recently, Weibo released its financial report for the first quarter of 2025, and its stock price rose 4.51% in pre-market trading, which is not only due to the solid performance of the advertising business, but also reflects the capital market’s recognition of its AI layout.
Equally striking is Tencent, whose stock price continued to rise after WeChat accessed the DeepSeek model, once hitting a new high in nearly three years.
These market reactions fully show that AI search is becoming a new engine to drive the value growth of Internet enterprises.
The market’s recognition and high expectations for AI search are actually also because of the new imagination it brings, whether it is in terms of content ecological upgrades, user experience optimization or platform user stickiness, AI search has played a blessing role, so as to make the platform’s traffic value higher.
This year, it is still a critical period for the development of AI search, and the competition is far from the end, how to build user trust in the rapid iteration of products, scenarios and technologies, and convert short-term traffic into long-term user retention will be a common problem faced by players.
In this race around the content ecosystem, the future winners will be those platforms that can deeply integrate technological dividends with ecological uniqueness.
