Around 40% of visitors arrive at adtechexplained.com through a Google search, and 20% of subscribers to Ad Tech Explained learned about the newsletter through Google.

Unsurprisingly, a 40% decline in display advertising revenue and a 20% decline in newsletter subscriptions would ensue if the pervasive adoption of these AI tools caused that search traffic to vanish.

Expanding this beyond a rudimentary ad technology website to encompass all publishers presents the Internet with a potential existential crisis. One in which AI response engines could effectively reduce advertising revenue and, as a result, publisher traffic by nearly half.

In light of users adopting generative AI search, Gartner recently predicted that search engine volume will decline by 25% by 2026, following a similar forecast that organic search traffic would decline by 50% by 2028. That is a substantial amount of potential traffic and revenue that may vanish in the near future.

What will the future hold for the web in light of this shift in user behaviour? How can publishing houses endure? Further, define an answer engine. We will investigate these concerns and more in this article.

Define an answer engine?

An answer engine is a chat application propelled by artificial intelligence that provides responses to any inquiry and has the potential to eventually supplant conventional search engines such as Google. They generate structured natural language responses by synthesizing and summarizing information gleaned from web navigation in order to provide direct response.

The term “Answer Engine” is one of my favourites because it precisely designates a particular AI conversation tool function as a direct competitor to search engines. Although rumours of a dedicated search feature for ChatGPT surfaced just this week, tools such as Perplexity that compete directly with search engines like Google more precisely suit the term, despite the fact that ChatGPT itself can be classified as an answer engine.

To observe the operation of an answer engine, please refer to this Perplexity result. “What is the ad selection API?” I inquired. Additionally, Ad Tech Explained was cited as the primary source.

It follows that the majority of users are not required to visit Ad Tech Explained, which is nice.

The following illustrations from The Browser Company’s Arc Search application, an additional answer engine, may provide insight into the manner in which the majority of users will interact with the Internet in the future.

My inquiry to Arc Search was, “What is Google PAIR?” It is noteworthy that my PAIR explainer is cited as the initial content source.

Arc search generates a completely original piece of content through the utilization of references to my content. The extent to which an individual is interested in learning about PAIR may determine whether or not they visit Ad Tech Explained, thereby nullifying my opportunity to acquire a new user or generate advertising revenue.

It is noteworthy to mention that the products of both Perplexity and Arc Search are powered by OpenAI models.

Headwinds for Publishers in the Age of AI

Answer engines are not the sole obstacles confronting publishers; they merely present one peril amidst the myriad of challenges that are relentlessly advancing against web publishers:

Google is imposing the Privacy Sandbox on publishers, compelling them to fundamentally restructure their advertising revenue generation strategies from passing web users (apparently oblivious to the decline in revenue they generate for the Google Network themselves).

Forbes and other made-for-advertising websites are stealing advertising revenue from publishers who adhere to the rules.

AI-generated spam is inundating the Internet, hijacking valuable search rankings and ad revenue. In response, Apple is developing Safari, the most popular mobile web browser in the United States, which will include a web eraser/ad blocker.

Time and focus continue to be dominated by walled gardens, leaving little room for the open web.

The ad-supported open web, which provides a fertile ground for a wide variety of publishers, is probably experiencing a severe contraction. What, however, does this imply for the future?

The future of a website powered by AI

Thus, allow me to summarize. In anticipation of the imminent siphoning of publishers’ traffic by AI tools, major technology conglomerates are stifling their efforts to monetize the remaining users. From here, where exactly do web publishers proceed? Few options exist beyond that:

• Abandon existence

The most sombre truth is that numerous publishers, their operations and clientele decreasing, will be forced to cease operations.

A significant proportion of the “open” Internet is comprised of listicles, endless slideshows, and low-quality content so densely crammed with advertisements that it is difficult to breathe. While this may appear favourable to some, it is not.

The actual tragedy is the possibility that access to journalism may be restricted. Advertising, whether one wishes to accept it or not, finances journalism and establishes a framework in which impartial, well-researched information is available to all. In the absence of advertising, subscriptions represent the only viable alternative business model for journalistic publications.

• Consider subscriptions

Those who do not have a subscription to reputable publications such as The New York Times, The Wall Street Journal, or Business Insider have likely encountered the aggravation of accessing a website featuring a captivating headline only to be greeted by a prohibitive payment barrier.

A couple of years ago, I examined the ethical implications of advertising versus subscription business structures; if you wish to delve more deeply into this subject, then do so. You will learn in that article the regrettable consequence of ad-supported content becoming unsustainable as a business model: access to information is restricted to individuals with financial means.

In light of the aforementioned challenges, could one fault a publisher for seeking sanctuary behind a secure paywall comprised of recurring and sustainable subscription revenue?

However, it is evident that not all publishers offer sufficient value to justify a subscription, and in this era of subscription fatigue, the number of publishers on the web is unsustainable due to the lack of market demand.

The burden of a deluge of subscription services is being felt by consumers. Do you believe that the majority of Netflix users would cancel in order to peruse the news?

• Increase the advertising value that users provide

The positive news is that the remaining inventory would become more valuable as a result of a decline in ad supply caused by publishers folding. Web publishers that manage to endure the threat of extinction are able to employ effective strategies that enable them to flourish despite the new limitations.

These approaches consist of further advocating for users to refrain from disabling their ad blockers, expanding support for Google Privacy Sandbox, and exploring alternative ID solutions.

Publishers must educate advertisers to redirect their spending away from MFA (Made for Advertising) sites and toward legitimate publishers in order to maximize the remaining traffic. Price will perpetually be under downward pressure if MFA sites continue to increase the supply of available ads; this is a lethal combination in an era of declining traffic.

• Grant Content Licenses to AI Firms

The strategy of “if you cannot beat them, join them.” How can publishers generate revenue from the content that AI tools scrape?

A license for Reddit’s content to be utilized by Google in training its artificial intelligence models was negotiated for $60 million. OpenAI entered into licensing agreements with the Financial Times, Axel Springer, and others to provide content for its models.

Large media companies are likely to continue to enter into similar agreements, as an AI model will benefit from a continuous pipeline and inventory of high-quality content that serves as training data.

This may result in a competition among AI companies to secure exclusive content sources, potentially leading to the stratification of information access and the introduction of political or moral biases influenced by the sources with which they conduct business. Nevertheless, prior to deciding, certain AI companies and publishers might be awaiting the resolution of a number of pending legal disputes.

Microsoft and OpenAI were sued for copyright infringement by The New York Times, which alleged that OpenAI used the plaintiff’s content to train its models. The lawsuit was filed against OpenAI after the New York Times failed to negotiate an appropriate licensing agreement.

This peril has not deterred AI companies from developing their tools at an accelerated rate. Microsoft has additionally made a commitment to users of its generative AI tool, Copilot, that in the event that a content proprietor contests a user’s Copilot creations on copyright grounds, they will bear full responsibility and legal liabilities.

AI companies are progressively becoming more cognizant of the legal ramifications associated with consuming content from any source. This is exemplified by OpenAI’s CTO’s incoherent response when queried about the utilization of YouTube or Instagram videos as training material for its video generation tool, Sora.

OpenAI argues that their models do not retain data in the manner of a database; instead, they generate new information by recognizing relationships in data. This reasoning suggests that they ought to be exempt from copyright claims, which is a weak argument. Open AI further employs an analogy of the same low quality: ChatGPT is comparable to an educator who has retained a wealth of prior knowledge and is capable of elucidating concepts due to her comprehension of their interrelationships; however, she does not retain the information in her memory.

There are not many educators I am aware of who possess omniscience and can respond instantaneously to any conceivable query posed simultaneously by millions of people.

Although OpenAI continues to secure licensing agreements, the majority of publishers fail to generate sufficient content to merit such agreements. As a result, these publishers are dependent on answer engines that construct the remuneration structure for all other entities.

Answers of Engine Optimization

An emerging concept that is becoming increasingly apparent to me is “Answer Engine Optimization,” or AEO.

The process of optimizing and structuring content in order to enhance its probability of being utilized as source material for answer engines is referred to as “Answer Engine Optimization.” AEO functions in the same way that SEO does for search engines.

Despite the primary objective of these tools being to deliver precise responses, they frequently cite the sources from which they obtained the information. Consequently, although the potential for referral traffic is considerably diminished compared to that of conventional search engines, there remains a chance to showcase your brand and potentially attract a visitor to your website.

Furthermore, the outcome of the litigation discussed in the previous section may potentially compel search engines to establish revenue-sharing agreements with the information sources.

Although Perplexity does provide a paid “Pro” service, in order to attain the same level of prominence as modern search engines, answer engines require a free alternative; this is where advertising could be advantageous. However, does any advertising business model resemble sponsored links on Google queries in a transparent and viable way?

In a recent episode of the Hard Fork podcast, Aravind Srinivas, CEO of Perplexity, presents this inquiry (59:00) to which he himself is unable to provide an answer. Nevertheless, he does suggest that Google’s substantial search advertising revenue line may prevent the company from promptly implementing these features (classic innovator’s dilemma).

When queried about the rationale behind publishers not being alarmed by the potential impact of search engines and Perplexity on publisher monetization potential, Srinivas failed to furnish a definitive explanation but expressed a readiness to collaborate with publishers in addressing the issue.

Publishers can only anticipate that legal rulings will ultimately mandate answer engines to compensate content sources; however, the manner in which these engines will generate revenue from advertising or remunerate publishers is yet to be established.