The New Information Ecosystem: From Search Results to AI Answers

The digital landscape is undergoing its most significant transformation since the advent of social media. The established paradigm of search-driven discovery, which has dictated digital marketing strategy for two decades, is being fundamentally challenged. A new ecosystem is emerging, one where users seek direct answers from conversational AI rather than a list of links to click. This report provides a strategic analysis of this shift, using Metehan Yesilyurt’s “AI Share Button” concept as a lens through which to examine the new rules of brand visibility and the nascent field of Large Language Model (LLM) optimisation.

The Great Traffic Re-Routing: The Decline of the Click and the Rise of the Answer

The foundational contract between search engines and content creators is changing. For years, the goal of Search Engine Optimisation (SEO) was clear: achieve a high ranking on the Search Engine Results Page (SERP) to win the user’s click. Recent data indicates this model is eroding at an accelerated pace.

A landmark 2024 analysis revealed that approximately 60% of all Google searches now conclude without a single click to a third-party website. This phenomenon, termed “zero-click searches”, signals a profound shift in user behaviour. The trend is even more acute on mobile devices, where the zero-click rate has surged to between 71% and 77%. This is not a cyclical downturn but a structural re-architecting of the SERP. Google is evolving from a simple directory of web links into a comprehensive answer destination. It increasingly satisfies user queries directly through on-page features like Featured Snippets, Knowledge Panels, and, most disruptively, AI Overviews. The direct consequence for businesses is a tangible loss of traffic, with some content marketers reporting organic visit declines of 20-30% despite maintaining stable keyword rankings.

The primary accelerator of this trend is Google’s AI Overviews, which were widely rolled out in May 2024. Their integration into the SERP has been aggressive. In January 2025, AI Overviews appeared in 6.49% of search queries; by March 2025, that figure had more than doubled to 13.14%. Industry estimates suggest these AI-generated summaries are responsible for a further reduction in organic click-through rates (CTR) of between 20% and 40%. The impact is concentrated at the top of the marketing funnel, as 88.1% of queries that trigger an AI Overview are informational in nature, directly competing with the blog posts, guides, and articles that form the bedrock of many content marketing strategies.

While this data has fuelled popular narratives about the “death of SEO,” such claims are largely unsupported. Google’s total query volume continues to expand, dwarfing that of its AI-native competitors. The crucial distinction is that the function of search is evolving. A growing portion of user journeys that begin on Google are now being redirected within its own ecosystem. Clicks to Google-owned properties like YouTube and Maps have increased, while clicks to external organic results have fallen.

This evidence points towards a necessary evolution in strategy. The objective is no longer simply to optimise for a clickable link. Instead, the focus must shift towards dominating the entire SERP. When a brand’s information is featured in a zero-click element like an AI Overview, it gains significant authority; one study found that nearly 70% of users place more trust in brands featured in these prominent positions. The strategic goal therefore transforms from “Search Engine Optimisation” to “SERP Dominance.” The value is captured not just through traffic, but through brand visibility and authority established directly within the search environment. This new form of brand awareness, sometimes called Answer Engine Optimisation (AEO), happens before a click ever occurs. It is this same challenge of achieving visibility within an answer-first environment that strategies targeting LLMs aim to solve.

The Ascent of the Answer Engines: A Market Analysis of the New Discovery Platforms

Parallel to the evolution of Google, a new class of “discovery engines” has emerged and rapidly achieved mainstream adoption. These LLM-powered platforms are not merely search adjuncts; they are high-traffic destinations where millions of users now begin their information-seeking journeys. For brands, understanding this fragmented landscape is the first step towards developing a multi-channel discovery strategy.

ChatGPT, developed by OpenAI, remains the clear market leader. As of April 2025, it boasted an estimated 800 million weekly active users and recorded 4.5 billion website visits in March 2025 alone. Processing over a billion queries daily, its scale is immense, with 92% of Fortune 500 companies reportedly using OpenAI’s products. Its user base is predominantly young, with 54.85% of users aged between 18 and 34.

Perplexity AI has carved out a distinct niche by positioning itself as an “answer engine” rather than a conversational chatbot. This focus on sourced, accurate information has fuelled explosive growth. By May 2025, the platform had reached 15 million active monthly users and 153 million monthly visits, representing a 191.9% year-over-year increase. Its query volume is reportedly growing at a rate of over 20% month-over-month, and it commands high user engagement, with an average visit duration of 23 minutes.

Claude, from Anthropic, is another formidable competitor. Backed by substantial investments from Amazon and Google, the company is projected to generate $2.2 billion in revenue in 2025. The platform attracted 16 million unique website visitors in January 2025, having peaked at 18.8 million in late 2024. Like ChatGPT, its user base skews young, with 51.88% of its audience aged 18-24.

Grok, from xAI, is a newer but rapidly ascending player, distinguished by its tight integration with the real-time data stream of X (formerly Twitter). Following the release of its Grok 3 model, the platform’s usage surged, reaching 35.1 million monthly active users and 141.9 million monthly visits by March 2025.

The data reveals a critical fragmentation of information-seeking behaviour. Users are not abandoning Google wholesale; indeed, 99% of those who use AI platforms also continue to use traditional search engines. Instead, they are assembling a “toolkit” of discovery sources, selecting the optimal tool for a specific task. A user might employ Google for a simple navigational query, turn to Perplexity for in-depth research requiring citations, and use ChatGPT for creative brainstorming or drafting text. This behaviour implies that a monolithic, Google-centric discovery strategy is no longer sufficient. To remain visible, brands must now develop a presence and a tailored strategy for each of these distinct answer ecosystems where their target audience is actively seeking information.

Under the Hood: How LLMs Find and Cite Information

To influence these new discovery engines, it is essential to understand the core technology that allows them to provide relevant, up-to-date answers: Retrieval-Augmented Generation (RAG). RAG is an AI framework that addresses the primary limitation of LLMs: their static, pre-trained knowledge. A model like ChatGPT, without augmentation, has a “knowledge cutoff” and cannot access information created after its training date. RAG solves this by connecting the LLM to external, live data sources, such as the internet or a proprietary company database, at the moment a query is made.

The RAG process can be understood in a few key steps:

1. User Query: A user submits a prompt to the LLM.
2. Retrieval: The RAG system converts the user’s query into a numerical vector. It then searches an external knowledge base (often a specialised vector database containing “chunks” of text from websites, documents, etc.) to find information that is semantically similar to the query.
3. Augmentation: The most relevant retrieved information is then packaged with the user’s original prompt. This creates an “augmented prompt” that provides the LLM with fresh, specific context.
4. Generation: The LLM receives this augmented prompt (original question + retrieved facts) and generates a response that is now grounded in the provided external data. This process dramatically improves factual accuracy, reduces the likelihood of “hallucinations,” and allows the model to answer questions about recent events.

For brands, RAG is the critical mechanism. It is the technical gateway that allows a brand’s public-facing content, such as a blog post or technical white paper, to be dynamically pulled into an LLM’s answer. A strategy like AI Share Buttons is designed to directly intervene in this process by proactively feeding a specific piece of content into the retrieval and augmentation step, thereby influencing the final generated output.

Deconstructing Metehan’s “AI Share Button” Growth Hack

Against this backdrop of shifting user behaviour and new technology, growth marketer Metehan Yesilyurt proposed a novel tactic designed to secure brand visibility within LLM environments. The “AI Share Button” is a user-activated widget on a website that enables a reader to push a piece of content directly into an LLM, packaged with a brand-defined prompt.

The CiteMET Framework Explained

The strategy is guided by the CiteMET framework, an acronym that outlines the four key objectives of the tactic:

• Cited: The primary goal is for the brand’s content to be explicitly named and linked as a source in the LLM’s generated response.
• Memorable: The interaction aims to create a persistent and positive brand association within the user’s personal chat history, potentially influencing the LLM’s future responses on related topics.
• Effective: The pre-engineered prompt that accompanies the content is designed to guide the LLM towards a specific, favourable output, such as a summary that highlights the brand’s key strengths.
• Trackable: The usage of the button (e.g., click-through rate) and its subsequent downstream effects (e.g., changes in branded search volume) should be measurable to gauge its impact.

In practice, this manifests as buttons placed under an article or on a product page with labels like “Summarise in ChatGPT,” “Analyse this data in Perplexity,” or “Ask Claude about this report”.

The Mechanism of Influence: Prompt Seeding and Context Injection

The core mechanic of the AI Share Button is context injection. When a user clicks the button, it triggers an action that populates the chosen LLM’s prompt box with two distinct elements:

1. The Content: The full text, or a significant, relevant portion, of the webpage the user is currently viewing.
2. The Prompt: A carefully crafted instruction, written by the brand, that directs the LLM’s analysis of the provided content.

For example, a button on a B2B software company’s blog post comparing its product to a competitor might generate the prompt: “Using the detailed comparison in the article provided below, create a table that highlights the unique advantages of Product X.”

This process is a form of proactive prompt engineering. It allows the brand to seize control of the narrative. Instead of leaving the user to formulate a vague or potentially biased query, the brand provides both the authoritative source material and the specific lens through which the AI should interpret it. This aligns with a modern marketing philosophy that prioritises shaping user intent over simply targeting keywords.

The Promised Land: Claimed Benefits and Strategic Goals

According to the strategy’s proponent, the successful implementation of AI Share Buttons can yield several strategic benefits that address the challenges of the new discovery landscape:

• Passive Citations in LLM Outputs: Earning direct mentions and, where possible, hyperlinks back to the brand’s domain from within AI-generated answers, creating a new form of referral.
• Persistent Brand Memory: Establishing a strong, positive association between the brand and a specific topic within a user’s private AI chat history, which serves as a valuable, long-term brand impression.
• Enhanced Perceived Authority: By being positioned as the primary source material for an AI’s authoritative answer, the brand’s credibility and expertise in its niche are reinforced.
• A Potential Bypass to Declining SEO Traffic: Creating a novel discovery channel that is not wholly dependent on the diminishing returns of traditional organic search rankings and click-through rates.
• Improved On-Site Engagement: Offering a novel, interactive tool for intellectually curious readers, which can increase session depth, time on page, and overall user engagement.

Critical Assessment: Hype, Reality, and Hidden Opportunity

While the concept of AI Share Buttons is intellectually appealing, its practical validity hinges on a rigorous assessment of its underlying claims. The effectiveness of this growth hack is not a simple yes or no; it is a complex function of LLM architecture, user behaviour, and the evolving definition of marketing ROI.

Are LLMs Credible Discovery Engines? A Reality Check

The first claim to test is whether LLMs represent a significant enough discovery channel to warrant dedicated optimisation efforts. While the query volume on platforms like ChatGPT and Perplexity is still a fraction of Google’s, this comparison can be misleading. The nature of the queries is often fundamentally different. LLM queries tend to be more complex, conversational, and geared towards problem-solving and in-depth research, whereas Google continues to dominate simple navigational and transactional searches.

Therefore, LLMs are not a wholesale replacement for Google but have emerged as a powerful complementary channel. They represent a new “discovery surface” that is particularly valuable for brands in knowledge-intensive sectors like B2B SaaS, finance, healthcare, and technology. For these companies, the audience using LLMs for research is a high-value demographic. The strategy is thus not about abandoning SEO, but about expanding the definition of “discovery” to include these new, highly engaged answer ecosystems.

The Citation Equation: Can AI Share Buttons Truly Influence LLM Outputs?

The core promise of the AI Share Button is its ability to generate citations. However, its efficacy is not uniform across all platforms. The outcome is dictated entirely by the specific LLM’s underlying architecture and its philosophy on sourcing information.

Perplexity AI, the self-proclaimed “answer engine,” is architecturally primed for this strategy to succeed. Its entire value proposition is built on providing transparent, sourced answers. Perplexity actively crawls the web with its own bot and appears to prioritise sources it deems authoritative for a given subject. A recent study of 30 million AI citations found that Perplexity’s most common source is Reddit (46.7%), followed by review platforms and professional content hubs. This indicates a preference for community-driven discussions and expert-level content. By using an AI Share Button to directly inject a high-quality article into a Perplexity prompt, a brand is providing the engine with precisely what it is designed to consume: a primary source document to analyse and cite. In this context, the likelihood of earning a direct citation is high.

ChatGPT, by contrast, is a “conversationalist” first and a research tool second. Its primary function is to generate text based on patterns in its training data, not to cite external sources. By default, it does not provide citations and is known to “hallucinate” or invent information. The same citation study found that ChatGPT’s top source is overwhelmingly Wikipedia (47.9%). This suggests that when it does retrieve external information, it defaults to a single, vast, and generally trusted corpus rather than a diverse set of web documents. For ChatGPT, an AI Share Button is therefore less likely to result in a direct, hyperlinked citation. Its value lies elsewhere. The button is highly effective at influencing the content of the generated summary. It forces ChatGPT to ground its response in the brand’s provided text, effectively controlling the narrative for that specific user interaction. The strategic win is not a backlink, but narrative control.

This architectural divergence reveals that the value proposition of the AI Share Button strategy is platform-dependent.

• For a platform like Perplexity, the goal is to become a citable source of fact.
• For a platform like ChatGPT, the goal is to become the basis for a narrative.

This distinction has direct practical implications. The pre-engineered prompts themselves must be tailored to the platform and the desired outcome. A prompt for Perplexity might be, “Analyse the data in the provided article and cite the source in your response.” A prompt for ChatGPT might be, “Based on the text below, write a summary explaining the unique benefits of this solution for a non-technical audience.”

(Source: Profound, analysis of 30 million citations from August 2024 – June 2025)

Rethinking ROI: Is This a Viable Growth Hack for Traffic?

Judged by traditional metrics, AI Share Buttons are unlikely to be a significant driver of direct, attributable referral traffic. These are not high-volume channels in the same way as organic search or large-scale social media campaigns. Attempting to justify this tactic based on direct clicks will likely lead to disappointment.

The true return on investment is more nuanced and lies in building brand equity within the “dark funnel” of AI interactions. This concept is analogous to “dark social,” where valuable brand advocacy and word-of-mouth recommendations occur in private channels like WhatsApp, Slack, or email, which are invisible to standard analytics tools. Every time a user’s query is answered using a brand’s content, a positive and authoritative brand touchpoint is created. The AI Share Button is a tool to intentionally generate these valuable, albeit hard-to-measure, touchpoints.

The impact of this activity can be measured through proxy metrics. A successful strategy should lead to an observable increase in branded search volume and direct website traffic over time. As more users are introduced to a brand through AI-generated answers, they are more likely to subsequently search for the brand by name or navigate directly to its website. This strategy should be viewed as a complement to, not a replacement for, established channels like SEO and email marketing. It is an emerging channel for early adopters to build authority in a new competitive arena.

Prompt Engineering as Narrative Control

The most potent element of the AI Share Button strategy is the pre-seeded prompt. It grants the brand a unique opportunity to frame the conversation and guide the AI’s interpretation of its content. This moves the brand from a passive subject of the AI’s analysis to an active participant in shaping the output.

Practical use cases are varied and powerful:

• For a blog post: “Using the article below, create a bulleted list of the top five takeaways for a busy executive, focusing on actionable advice.”
• For a product page: “Based on the following technical specifications, explain in simple terms how this product solves the problem of data integration for small businesses.”
• For a case study: “Summarise the key results from this case study into three quantifiable outcomes that would be relevant to a Chief Financial Officer.”

This power must be wielded responsibly. There is a fine line between guiding an AI to accurately interpret high-quality content and attempting to mislead users by forcing the generation of false or exaggerated claims. Any growth hacking technique that deceives users is detrimental in the long run. The ethical application of this strategy involves using prompts to enhance clarity and highlight genuine value, not to manipulate the AI into fabricating information.

A Practical Guide to Implementation and Measurement

Translating this analysis into action requires a disciplined, experimental approach. The following guide provides a framework for brands to test the AI Share Button strategy, positioning it as an exploratory tactic rather than a core pillar of the marketing mix.

Who Should Experiment and How?

This strategy is not universally applicable. It is best suited for organisations whose primary currency is knowledge and expertise.

• Ideal Company Profiles:
  • B2B SaaS Companies: Particularly those with complex products that require significant educational content, technical documentation, and thought leadership to explain their value.
  • Knowledge-Heavy Publishers: Brands operating in finance, technology, law, science, or other fields where authority and in-depth analysis are key differentiators.
  • E-commerce with Educational Content: Retailers selling high-consideration or technical products (e.g., high-end electronics, specialised equipment) that are supported by extensive guides, tutorials, and comparison articles.
• Best Pages for Testing:
  • High-Performing Evergreen Content: The most popular and authoritative articles that already attract significant organic traffic. These pages have proven value and are prime candidates for AI summarisation.
  • Cornerstone Pages and Pillar Content: Foundational, long-form guides that define key concepts in an industry. These are ideal for establishing a brand as a definitive source for AI models.
  • Technical Documentation and White Papers: Dense, factual content that is perfectly suited for AI-powered analysis and summarisation, a task many users seek to perform.
• Example Prompt Styles to Test:
  • For Summarisation: “Provide a concise summary of the key arguments in the article below.”
  • For Benefit Extraction: “Based on the following text, list the top 3 benefits of [Product Name] for.”
  • For Comparison: “Using the data in the provided text, create a comparison table between [Product A] and.”
  • For Data Analysis: “Analyse the financial data presented in the report below and identify the key trends.”

A Technical Guide to Implementation

Implementing AI Share Buttons can range from a simple, no-code solution to a more customised manual setup.

Using a URL Generator (The Easy Way)

For those who want to test the concept quickly without writing code, the creator of the strategy, Metehan Yesilyurt, provides a free AI Share URL Creator. This tool allows you to input your content’s URL and a custom prompt, and it will generate the specific share links for various LLMs.

Metehan’s AI Share URL Creator: https://metehan.ai/ai-share-url-creator.html

Manual Implementation with Code

For more control over the appearance and functionality, you can build the buttons directly into your website using HTML, CSS, and JavaScript.

1. Add Buttons to Your Content

Place these buttons in your article where you want them to appear. The `data-prompt` attribute holds the instruction for the AI.

2. HTML Structure Example

This is the basic HTML you would add to your post’s code.

<!-- AI Share Buttons -->
<div class="ai-share-buttons">
    <button class="ai-share-btn" data-llm="chatgpt" data-prompt="Summarise the key points of the following article for a business executive:">
        Summarise in ChatGPT
    </button>
    <button class="ai-share-btn" data-llm="perplexity" data-prompt="Analyse the following text and provide a list of citable facts:">
        Analyse in Perplexity
    </button>
    <button class="ai-share-btn" data-llm="claude" data-prompt="Based on the article below, what are three potential counter-arguments to the main thesis?">
        Discuss in Claude
    </button>
</div>

<!-- Your article content must be wrapped in an element with this ID -->
<article id="article-content">
    <p>This is the full text of your amazing, insightful article...</p>
</article>

3. JavaScript for Functionality

This script handles the logic. In WordPress, you can add this using a custom HTML block or a dedicated plugin for adding scripts.

document.addEventListener('DOMContentLoaded', () => {
    const shareButtonsContainer = document.querySelector('.article-container');

    if (shareButtonsContainer) {
        shareButtonsContainer.addEventListener('click', (event) => {
            if (event.target.matches('.ai-share-btn')) {
                const button = event.target;
                const llm = button.dataset.llm;
                const promptText = button.dataset.prompt;
                
                // Get the current page's URL instead of the full text
                const articleUrl = window.location.href;

                // Combine the brand's prompt and the article URL
                // This instructs the LLM to fetch and process the content from the link.
                const fullPrompt = `${promptText}\n\nArticle to process: ${articleUrl}`;

                const encodedPrompt = encodeURIComponent(fullPrompt);
                let shareUrl;

                switch (llm) {
                    case 'chatgpt':
                        shareUrl = `https://chat.openai.com/?q=${encodedPrompt}`;
                        break;
                    case 'perplexity':
                        shareUrl = `https://www.perplexity.ai/?q=${encodedPrompt}`;
                        break;
                    case 'claude':
                        shareUrl = `https://claude.ai/chats?q=${encodedPrompt}`;
                        break;
                    default:
                        console.error('Unknown LLM target:', llm);
                        return;
                }
                window.open(shareUrl, '_blank');
            }
        });
    }

    // Functionality for all "Copy" buttons
    const allCopyButtons = document.querySelectorAll('.copy-button');
    allCopyButtons.forEach(button => {
        button.addEventListener('click', () => {
            const pre = button.parentElement.querySelector('pre');
            const code = pre.querySelector('code');
            navigator.clipboard.writeText(code.innerText).then(() => {
                button.innerText = 'Copied!';
                button.classList.add('copied');
                setTimeout(() => {
                    button.innerText = 'Copy';
                    button.classList.remove('copied');
                }, 2000);
            }).catch(err => {
                console.error('Failed to copy text: ', err);
            });
        });
    });
});