The Giants' Covert Battle in AI Marketing

Author | Zhou Zhiyu

Editor | Zhang Xiaoling

As the market enters the era of stock, the golden age where one could blindly buy traffic to exchange for GMV is long gone. The anxiety of advertisers is visibly apparent: budgets are shrinking, but the demands for results have become almost harsh.

Beneath the frozen ground of traffic, a dark current is surging wildly.

In stark contrast to the sluggish market, the AI marketing sector is expanding at an astonishing pace. According to iResearch Consulting, the AI marketing market is expected to exceed 85 billion yuan by 2025, a year-on-year growth of 127%. A survey by consulting firm Capgemini shows that 79% of marketing technology investments are flowing into AI.

This sector has achieved exponential growth in just a few years, and tech giants, after reaching the peak of traffic, are attempting to reshape marketing into a money-printing machine using AI.

The atmosphere of competition has never been so intense. Alibaba Mama, Douyin Engine, Tencent Advertising, and Baidu Marketing have all placed their bets on AI. From using AI to tap into the dividends of short dramas and mini-games to optimizing recommendation algorithms, players in the field are showcasing a variety of skills.

In this chaotic battle, Baidu has chosen a more thorough path: it aims to change the way users interact with the internet.

Zhang Lihong, head of Baidu's commercial products, told Wall Street Insights that the core of today's advertising change lies in the increased choices available to users, and the diversity of user needs. Traditional hard ads are no longer appealing to users.

A noteworthy detail is that users now have higher expectations for search results. Previously, users would search for "Logitech mouse," but now they directly search for related jargon. Traditional keyword advertising often struggles to handle this vague, colloquial, and even emotional jargon, merely pushing a bunch of links containing the word "mouse" to users.

However, Baidu sees this as an opportunity. Zhao Xin, head of Baidu's merchant intelligence product, stated that with the "merchant intelligence" driven by large models, the current search engine can not only understand "jargon" but can also respond to users' deeper intentions like a knowledgeable sales assistant, thus completing the loop from "inquiry" to "order." This change in interaction marks a fundamental shift in marketing logic: internet advertising has finally evolved from static display to proactive service.

The impact of AI marketing technology transformation is most deeply felt by small and medium-sized enterprises.

Imagine a small law firm with only three to five people; they face a dilemma that is both real and helpless: competition for case sources is fierce, and many incoming consultation calls are often ineffective, leaving lawyers exhausted from answering repetitive questions like "Is it civil or criminal?" and "What is the amount involved?"

Zhao Xin mentioned that for these small law firms lacking online marketing experience, AI intelligence serves as the tireless "front desk," automatically filtering intentions and lowering the marketing threshold.

Now, the lawyer intelligence can complete the first round of case sorting and intention layering just like a real person, and even automatically assess the case value based on user responses. If it is a simple consultation, the intelligence provides direct answers; if it is a major case, the intelligence guides the client to leave contact information or even directly schedule a time with a lawyer In response to the difficulty of understanding dialects and the inconvenience of typing, Baidu has also launched a "voice intelligent agent." Zhang Lihong added that in B2B scenarios, many clients are both bosses and investors, and often struggle to understand consultation calls filled with dialects. The voice intelligent agent can accurately recognize dialects and transcribe them, perfectly addressing the pain points of "not understanding and being overwhelmed."

The feedback from data is the most direct. Zhang Lihong revealed a set of internal testing data: in traditional manual customer service scenarios, the speaking rate is usually only 5%-6%, while after being taken over by this emotionally stable merchant intelligent agent, the speaking rate doubled.

For small and micro business owners who are both "investors and bosses," this is not only a cost reduction but also an empowerment of capabilities. Zhang Lihong emphasized that Baidu's "minimalist deployment" strategy is designed to capture this group of small and medium clients with a high turnover rate—they do not need to understand complex large model parameters, and AI tools help them through the most challenging cold start period.

Looking back at the exploration over the past year, Zhang Lihong described the process of companies embracing AI as "from pain to joy."

"It was painful at first," she admitted. The forced integration of technology brought discomfort; technical staff were worried about having to completely reshape the development model, and product managers were concerned that large models might hallucinate, not knowing how to make previously static strategies dynamic. Business owners also did not trust this "black box," feeling that AI customer service was sometimes "intellectually disabled," only able to repeat clichés.

At that time, the strategy was to "cast a wide net," trying out the possibilities of AI in all aspects. But the turning point occurred at the moment the data started to flow. When the first hundred clients who took the plunge discovered that the funnel data had indeed improved and the cost of leads had really decreased, confidence was established.

Currently, more than 30,000 merchants are using Baidu's merchant intelligent agent every day. In addition to the intelligent agent, AI digital humans are another explosive point. IDC data indicates that the digital human market will experience astonishing growth by 2025, and at Baidu, this is not just a tool for e-commerce sales but has also become a magical tool for lead generation.

The business world in 2025 is undergoing a dramatic transformation from seeking dividends to honing internal capabilities.

Structural dividends are often just "small fortunes" in short cycles; once discovered, they can be instantly neutralized by countless competitors using the same AI tools. As the growth of buying traffic becomes history, business owners must begin to consider more fundamental questions: my budget not only needs to bring in a click but also to establish a deep operational connection.

In this sense, the "intelligent agent" promoted by Baidu Marketing is not just a marketing product; it is more like an interface for businesses to connect to the future commercial world.

Other players in the market are also making their own attempts. ByteDance's Douyin Engine remains an aggressive "traffic hunter," using AI to frantically mine new growth in short dramas and mini-games, attempting to combat growth anxiety with content dividends; Alibaba Mama is focused on "operational certainty," integrating e-commerce knowledge with reasoning capabilities through AI to accurately predict each user order; Tencent Ads chooses to build ecological high walls, using AI to integrate WeChat, video accounts, and other touchpoints, refining every inch of the vast private traffic pool In this new trillion-level track, the old kings may fall, but new legends are being written by those pioneers who are adept at using AI. This may be the most brutal yet fascinating moment in business history