With artificial intelligence revolutionizing traditional industries, a new concept is emerging to simplify our lives: no more sifting through irrelevant websites or spending hours researching. This is the future promised by agentic search. Perplexity, a rapidly growing startup, has challenged Google with its AI-powered search website since 2022. And that was just the beginning. They’ve teased the public with Comet, their next-gen browser that uses agentic search to deliver personalized results without endless queries.
One question should be considered: Will this mark the end of traditional search and signify a massive change in consumer behavior? AI adoption has been skyrocketing among Gen Z and the older generations in recent years. Leading tech companies such as Google and Microsoft understand well the power of large language models (LLMs) and are incorporating them into their existing services, such as Microsoft’s Copilot. Judging by those two trends, it is inevitable that a more intelligent form of searching, powered by the groundbreaking power of LLMs, will replace the traditional search.
Recent Change in User Behaviors
With AI agents available to provide tailored results without users submitting multiple queries, traditional search engines are quickly losing market share. The ease of access to personalized information is dragging early technology adopters away from the repetitive nature of traditional search. In everyday tasks like shopping and more sophisticated tasks like researching, users are finding more efficiency in using AI-powered searches. This shift in behaviors for a part of the user segment and user experience indicates a massive potential for agentic searches to invade the territories of other traditional search use cases.
While behaviors may take more apparent benefits to alter, simply having an interest has relatively lower costs. Everyday users have become more comfortable with AI tools in the past 2 years as companies have incorporated LLMs into their websites and products.
A recent survey noted that 41% of the US. Consumers are already interested in using AI-generated search results for online shopping, and this percentage is even higher for Gen Z, at 59%, who are actively using AI search summaries for at least half of their queries. With a close-to-majority portion of consumers understanding and appreciating the benefits of AI-generated results, this trend will likely continue to grow over the next few years until most consumers are entirely interested if not reliant on, AI tools.
Agentic search will likely ride on this trend and become a disruptive innovation that fundamentally changes the search landscape from the perspective of user behaviors. Such a shift in user behavior will force companies to modify their product offerings to sustain their business, ultimately leading to agentic search becoming the dominant way of searching.
Moves From Tech Giants
Tech giants like Google and Microsoft also anticipate this trend and have made several strategic moves in all layers of the LLM value chain. For example, both Google and Microsoft are already experimenting with AI-generated results within search queries with their in-house LLMs (Gemini and OpenAI’s GPT, respectively). Users now expect an AI summary when they search with those companies’ search engines, which is not far from what people can expect with agentic search. With leaders in the industry already modifying their existing products for experiments, if the outcome of those experiments is positive for the user and business sides, one can reasonably expect the rest of the industry to follow their leaders.
The user side is already showing positive trends with the massive adoption of AI technologies; the question is whether an enormous income provider for the search industry – digital advertisement – will gladly accept this change. Several leaders in traditional search, such as Google and Microsoft, and leaders in LLM and agentic search, such as OpenAI and Perplexity, are already discovering the potential to incorporate ads inside search results. These attempts have sparked a growing interest in companies needing advertisements. From 2023 to 2024, in just one year, the willingness to advertise alongside AI-generated content has increased from 33% to 57%.
With both the brands and consumers willing to accept the change with AI, this industry shift to agentic search shows more promise, and with momentum building up from both directions, not only should we expect startups to make the moves but also tech giants who have more financial resources, more established infrastructure and more significant incentives (both protecting their existing ecosystem and capturing more market shares from other giants).
Resistance From Both Consumers and Advertisers
While there is some momentum pushing the transition from traditional search to agentic search, many may expect considerable resistance from consumers and advertisers. For consumers, this means altering their deeply ingrained behavior and lifestyle associated with traditional google search, and for advertisers, this means modifying their established business models. The previous discussion appears overly optimistic when assessing both parties’ potential unwillingness to change their behavior.
Agentic search, which is heavily reliant on the capabilities of LLM, can often lose credibility due to hallucination, where the LLM makes up information. Due to this fundamental flaw, many may argue that traditional search still has an edge because it provides consumers who value transparency and credibility from the search results. Indeed, the benefit of adopting agentic search in contrast to traditional search may be vague for mainstream consumers, and having less control over the search process could even scare off many consumers.
Nonetheless, with the rise of social media in recent years, Gen Z has already been adopting a new way of searching called social searching, where they simply search on social media like TikTok and Instagram for information. Quickness and relevance are core factors pulling Gen Zs to change how they search for everyday details, and traditional searches often provide superficial information, requiring more time to find relevant information. Social searching offers users far less control since most information is created by other users on social media, which also has more potential for misinformation. Suppose young consumers are willing to switch to a search behavior with less control and credibility in exchange for quickness and relevancy. In that case, agentic search, which provides consumers with relevant references to actual websites, quotes, and more personalized information input, should also be attractive to those consumers.
LLM capabilities have also been improving rapidly. Claude—a long-time leader in creating LLMs—has gained massive trust from users with its 3.7 Sonnet model. Hallucination, a technical flaw arising from the training process of those models, can be significantly reduced with the model’s capabilities to reference accurate sources from the Internet, which is precisely what agentic search excels at. At this speed of improvement, one can expect equal if not more trustworthy results with agentic search. Ultimately, it appears that the concerns that consumers may hold will only be present in a short period when the leading companies pour enormous amounts of capital and talent into this new field.
While consumers may easily switch behaviors, the uncertainty of adopting a new business model for advertisers may be met with more resistance. Traditionally, advertisers use tangible metrics such as clicks and impressions to track the effectiveness of their ad campaign. Since most advertising companies rely on this fundamental system to communicate with external and internal stakeholders, operate, and essentially sustain their business, many experts may be concerned about this transition’s impact on the advertising industry. However, advertisers’ choice of advertisement channels mostly depends on the return on investment for any campaign. In other words, if agentic search can improve the effectiveness of those ads, helping them find the consumers who are the most interested, then it becomes a question of which search method will be more effective and which would attract more advertisers.
Meta, a social media tech giant, has proved that AI-powered ads can provide more returns. In 2024, Meta launched an AI-powered ad for advertisers, which delivered 22% higher returns than the average Meta ad. With Meta’s success, agentic search, which searches on behalf of the user’s request, can also potentially eliminate most inefficiencies in advertisements by presenting the users with a minimal amount of ads with high relevancy, translating into potential purchasing intents. In that sense, advertisers can simply deploy strategies that align with their existing infrastructure while earning a higher return, and for the consumers, there will be fewer annoying, irrelevant ads. No disruption will happen on the advertiser’s side except for receiving greater returns with an alternative channel.
With consumers’ trust issues and advertisers’ business model issues finding concrete solutions, overcoming the barriers of the transition seems to be a mere time issue. Many doubts about the actual feasibility of this transition are grounded in the fear of change, yet, as seen by recent events, consumers and advertisers are favoring the change and are expected to join the seamless transition readily.
What to Expect in the Next Few Years
So, what does this mean for consumers, advertisers, and the broad search industry? The wheel of transformation is silently yet steadily turning. Perhaps with a “ChatGPT” moment in agentic search, the world may witness another explosive change in consumer behavior, followed by countless companies chasing into the new market with new products.
That being said, the real question in this era of AI is, can we trust such a transition? Are we comfortable handing our controls to LLMs and permitting them to tell us what we want to see? In years from now, perhaps we will all be using a browser like Comet with our smartphones, entrusting those browsers to make the best decision for us during searches. But as we are handing more power to agents other than ourselves, how much independence do we still have, and will we still be able to maintain our independent thinking abilities?
LLMs have unlocked convenience for us, but also have the power to remove our thinking ability. If our sole source of information comes from those systems, do we still have the power to verify the information we are presented with? Many answers are still unclear at this point in history, but vigilance should be maintained at this tipping point in the search industry, along with the excitement of a massive transition.
