Positive Reviews: The Exploration of AI Search Ads, Difficult but Showing Positive Signals of Industry Breakthrough
The commercialization dilemma of AI search ads is essentially an exploratory experiment for “game-changers.” Although companies like Perplexity are currently facing issues such as low advertising revenue and executive departures, from an overall industry perspective, this exploration still emits multiple positive signals, pointing out possible breakthrough directions for the future of AI search.
Firstly, the “precision” potential of AI search ads has begun to emerge, injecting new value into the traditional advertising model. Traditional search ads rely on keyword matching by users, and ad spaces are presented in the form of “information lists,” requiring users to actively filter. In contrast, AI search can understand user intentions more deeply through natural language interaction and embed ads into the “answer generation” scenario. Microsoft’s practical data confirms this – the ad click – through rate of its AI product Copilot is 73% higher than that of traditional search, and the conversion rate has increased by 16%, especially in scenarios with clear intentions such as shopping, where the effect is more significant. This shows that although AI search ads are “small in quantity,” they are “high in quality” and can connect user needs with commercial supply more efficiently. This advantage of “precise conversion” may become the core competitiveness of future advertising models and even reshape the “effect evaluation system” of the advertising industry.
Secondly, the industry’s emphasis on user experience lays the foundation for a long – term trust relationship. Different from the rough model of traditional search where “ads are information,” players in the AI search field generally take a cautious attitude towards ad integration. For example, OpenAI has clearly regarded advertising as a “last resort,” emphasizing the need to design “very carefully and tastefully.” When Perplexity tried to embed sponsored links, it did not forcefully squeeze the user’s conversation space. This “user – first” strategy essentially protects the core value of AI search – users’ trust in AI as a “reliable advisor.” Once trust is established, the acceptance and conversion potential of ads will increase significantly. As a former Google executive said, “Users are willing to pay for trustworthy advice. If ads can become part of ‘useful advice,’ their value will far exceed simple traffic exposure.”
Thirdly, the strategic layouts of leading enterprises show the industry’s ambition and provide room for imagination for ecological expansion. Perplexity’s “bold bet” to acquire the Chrome browser for $34.5 billion, although regarded by the outside world as “a publicity stunt taking advantage of the anti – monopoly trend,” has a clear underlying logic – by controlling the browser entrance with a 70% global market share, it can quickly obtain 3 billion user traffic to make up for its own shortage in search volume. This strategy of “competing for traffic entrances” is similar to Google’s path of consolidating its search hegemony by bundling browsers back then, reflecting the importance that AI search players attach to the “ecological niche.” In addition, Microsoft’s integration of AI technology into Bing and Google’s exploration of Search Generative Experience (SGE) all indicate that the giants are building a more complex AI search ecosystem through the integration of “technology + scenarios,” creating more possible “touchpoints” for ad monetization.
Finally, the exploration of diversified monetization paths reduces the reliance on a single advertising model. Perplexity’s annualized revenue exceeds $100 million, mainly coming from subscriptions (Comet Plus) and API services. OpenAI is expected to earn $12.7 billion through subscriptions this year, far exceeding the possible short – term contribution of advertising. This combined model of “subscription + API + advertising” not only relieves the pressure of ad monetization but also provides enterprises with more stable cash flow. For example, by sharing subscription revenue with the media (such as an 80% share of Comet Plus), Perplexity not only alleviates copyright disputes but also converts subscription users into long – term “content – paying” customers, buying more time for trial – and – error in advertising business.
Negative Reviews: The Commercialization Dilemma of AI Search Ads Exposes Deep – seated Contradictions in the Industry’s Underlying Logic
Despite the positive exploration, the commercialization dilemma of AI search ads has become a “bottleneck” problem in the industry. From Perplexity’s quarterly advertising revenue of $20,000, to Google’s “glue pizza” fiasco in AI search, and OpenAI’s hesitant attitude towards advertising, behind these is the conflict between the underlying logic of AI search and the traditional advertising model, as well as the difficult balance that enterprises need to strike between “monetization” and “experience.”
Firstly, the “answer generation” feature of AI search naturally compresses the physical space for ads. Traditional search pages can display more than 10 ad links simultaneously, and users actively click through keyword filtering. In contrast, the interaction form of AI search is “dialogue – based,” and each query usually generates a summary answer. Ads can only be embedded at the end of the answer in the form of “sponsored links” or “recommended services,” with extremely limited positions. For example, in the pilot, Perplexity can only embed a small number of sponsored links in the Q&A interface, and users’ attention is concentrated on the answer itself, greatly diluting the “exposure opportunities” of ads. This “space limitation” directly leads to the eCPM (earnings per thousand impressions) of ads being much lower than that of traditional search, making it difficult to cover the high computing power costs of AI search (Google admits that the cost per query of AI search is higher than that of traditional search).
Secondly, users’ expectation of AI “neutrality” forms a fundamental contradiction with the “interest association” of ads. The core need of users when using AI search is to “obtain reliable answers.” Once they perceive that the answers are manipulated by advertisers, trust will collapse instantly. For example, Google’s AI search once recommended wrong suggestions such as “using glue on pizza to prevent cheese from sliding off” and “mixing bleach with white vinegar to clean the washing machine.” Although not directly intervened by ads, it has already triggered users’ doubts about the reliability of AI – if advertisers further influence the answers, the consequences would be even more unimaginable. Therefore, even when AI search embeds ads, it is necessary to clearly mark them as “sponsored content,” which in turn reduces the user click – through rate (Microsoft’s data shows that the click – through rate of clearly marked ads is 40% lower than that of implicit ads). This dilemma between “trust and monetization” reduces advertisers’ willingness to invest, and pilot cooperation projects of companies like Perplexity (such as with TurboTax and Whole Foods) are difficult to scale up.
Thirdly, copyright disputes and legal costs further squeeze the profit margin of ad monetization. Perplexity has spent millions of dollars on legal lawsuits in the past year. The collective lawsuits from media such as The New York Times and Nikkei are essentially protests against AI search’s “free use of content.” Although Perplexity has proposed a compromise plan of “sharing 80% of subscription revenue,” the media is more concerned about “advertising revenue sharing” – after all, the scale of subscription users is limited, and advertising is the major source of long – term monetization. If the advertising revenue of AI search grows in the future, the media may demand a higher sharing ratio (for example, the sharing ratio between traditional search engines and content providers is usually between 15% – 30%), which will directly erode the advertising profits of enterprises. For example, assuming that Perplexity’s advertising revenue reaches $10 million, if 20% is shared with the media, the actual revenue of the enterprise will only be $8 million, while the computing power and R & D costs of AI search may reach tens of millions, further squeezing the profit margin.
Finally, the short – term limitation of the industry’s advertising market scale is difficult to support enterprises’ high valuations and expansion ambitions. According to eMarketer data, the expenditure on AI search ads in the United States in 2024 is only $1 billion, and it is expected to reach $26 billion in 2029, accounting for only 13% of the overall search advertising market. Perplexity’s current valuation is $18 billion. If it only relies on advertising, it needs to increase its advertising revenue to billions of dollars in the next few years, which is almost impossible given the limited market scale. More importantly, the “money – burning” speed of AI search far exceeds that of traditional Internet companies – OpenAI is expected to achieve positive cash flow only in 2029. If Perplexity cannot quickly find monetization paths other than advertising (such as Agent services and enterprise – level APIs), its high valuation will be unsustainable.
Suggestions for Entrepreneurs: Find a Balance between “Monetization” and “Experience” and Build a Diversified Growth Engine
The advertising dilemma of AI search is essentially a collision between technological innovation and business logic. For entrepreneurs, they need to break out of the thinking of “copying traditional search ads” and build a differentiated monetization strategy based on the characteristics of AI. The following are specific suggestions:
- Optimize ad scenario design with “precise conversion” as the core: Abandon the “scatter – gun” ad space layout and focus on high – intention user scenarios (such as shopping, travel, and local services). Through AI, deeply understand user needs and transform ads into part of the “solution.” For example, when a user asks for “recommendations for a weekend parent – child trip in Beijing,” a “ticket + hotel” package link in cooperation with scenic spots can be embedded and marked as “high – quality service verified by user reviews,” which not only maintains the neutrality of the answer but also improves the conversion rate. Referring to Microsoft Copilot’s experience, focus on the dual indicators of “click – through rate + conversion rate” rather than simply pursuing the number of ad spaces.
- Build a diversified revenue structure of “subscription + advertising + service”: Avoid over – relying on advertising. Lock in high – sticky users through subscriptions (such as Perplexity’s Comet Plus), obtain stable B – side revenue through API services (such as providing AI search interfaces to enterprises), and then use advertising as a “supplementary monetization” method. For example, a small number of precise ads can be shown to free users, while “ad – free + in – depth services” (such as customized reports and priority customer service) can be provided to subscribed users, which not only increases users’ willingness to pay but also reduces the impact of ads on the experience.
- Proactively solve copyright issues and build a content cooperation ecosystem: Establish a “revenue – sharing” mechanism with the media and content providers instead of passively dealing with lawsuits. For example, in addition to sharing subscription revenue, 10% – 15% of advertising revenue can be shared with content providers, and the “content source” should be clearly marked, which not only alleviates legal risks but also enhances users’ perception of the credibility of answers. Perplexity’s “80% subscription sharing” can be used as a reference, but it needs to be extended to the advertising field, and a collaborative model of “content – advertising” should be jointly designed with the media.
- Explore the Agent model and shift from “selling attention” to “selling results”: The ultimate value of AI search is to “help users complete tasks” rather than just provide information. Entrepreneurs can develop “task – oriented Agents” (such as booking hotels, buying air tickets, and making service appointments) and obtain commissions or shares by facilitating actual transactions. For example, when a user asks for “recommendations for a cost – effective hotel in Shanghai,” the AI can directly call the interface of the cooperative platform, display hotel details, and provide “one – click booking,” and charge a 3% – 5% commission. This “result – oriented” monetization model not only avoids the interference of ads on the experience but also creates higher value per user.
- Control computing power costs and optimize technical efficiency: The high computing power cost of AI search is the core obstacle to ad monetization. Entrepreneurs need to reduce the cost per query through model optimization (such as parameter – efficient fine – tuning and inference acceleration) and computing power resource scheduling (such as using idle GPUs and hybrid – cloud deployment). For example, use a lightweight model for simple questions (such as “today’s weather”) and a large model for complex questions (such as “financial analysis”) to dynamically allocate computing power resources. At the same time, cooperate with cloud service providers to obtain a lower computing power procurement price and improve the ratio of advertising revenue to cost.
The advertising dilemma of AI search is both a challenge and an opportunity. Entrepreneurs need to break out of the traditional search thinking framework, focus on user needs, and combine the “understanding, generating, and executing” capabilities of AI to build a business model of “experience – first, diversified monetization.” Only in this way can they gain a foothold in the wave of AI search.
- Startup Commentary”Say Goodbye to Ineffective Busyness: First, Expose Your False Sense of Urgency”
- Chain Exploration”Is the End of Stablecoins a Public Blockchain? The New Attempts of the Big Three”
- Startup Commentary”Real entrepreneurial stories are completely different from what you see in big companies.”
- Chain Exploration”Musk’s Heavy Bet: Is the Next Disruptive AI Trend on the Horizon?”
- Startup Commentary”Stablecoins: The Last Straw That Broke the “North American Alipay””
