Tag: AI Search

  • How Brave Search Rankings Boost Claude’s AI Visibility

    How Brave Search Rankings Boost Claude’s AI Visibility

    I’ve discovered something intriguing about Claude’s reliance on Brave Search rankings. Based on insights shared by Jonathan Clark during a Profound session on Zero Click, it seems that Claude frequently taps into Brave’s search results, particularly when dealing with recency, ranking, or comparison prompts.

    Clark, who is the managing partner at Moving Traffic Media, emphasized a key point from the session: Claude doesn’t rearrange search results but instead incorporates Brave’s top 10 search results directly into its answers.

    Claude’s web searches are selective. In fact, I learned that Claude uses web search in only 36.6% of cases compared to about 90% for ChatGPT, as per Clark’s observation.

    Claude is triggered to search most often by prompts that signal current trends, rankings, location, or comparisons. For example, queries like “best XYZ” caused a search 81% of the time. Ranking focus prompts had a search rate of 67%.

    Location prompts initiated searches 55% of the time, while comparison prompts such as “X vs. Y” led to searches 51% of the time.

    Brave rankings are crucial. Another interesting point is that Claude’s answers only matched ChatGPT’s citations in 8% of cases for the same queries, according to Clark.

    Claude’s results showed a 64% overlap with Google rankings. This indicates that Google-focused SEO strategies might be more effective for Claude than efforts targeted at boosting visibility in ChatGPT.

    The analysis also highlights the significance of tracking Brave search rankings. Clark mentioned that Claude relies on Brave, and achieving good rankings in Brave provides us with measurable insights.

    Memory in prompts. I found it interesting that prompts like “how does,” “what is,” and “steps to” are less likely to prompt Claude to conduct a web search. Without searching, Claude cannot cite online sources.

    According to Clark, Claude searches most often for prompts with keywords like “best,” “top,” or comparative phrases.

    The pattern of years in queries. Clark noted that there are consistent patterns that might simplify testing with Claude:

    One noticeable trait is Claude’s query fan-outs, which consistently produced the same results 65% of the time across users.

    These fan-outs frequently involve years, suggesting that titles featuring the current year might be advantageous in Claude-initiated searches, especially for queries driven by ranking and recency.

    Why this matters to us. It appears that Claude’s visibility hinges more on the rankings within the search results it utilizes. Clark suggests Claude might be one of the most amendable AI answer engines due to its consistent search patterns closely tied to measurable rankings.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • Unlocking AI Search and Ads: Insights from Ginny Marvin

    Unlocking AI Search and Ads: Insights from Ginny Marvin

    After Google Marketing Live, I’m still left with a lot of questions, and I’m sure I’m not the only one. Thankfully, Ginny Marvin, Google Ads Liaison, joined a comprehensive Q&A with Julie Bacchini and the PPC Chat community to tackle big topics like AI Max, AI Search ads, first-party data, and more.

    The discussion was enlightening, bringing clarity to AI Search eligibility, reporting challenges, and Google’s increasing focus on data quality.

    AI Max: Not a Must-Have for AI Search Ads

    A major revelation was that AI Max isn’t required for participating in AI-driven search experiences. This surprised many of us, as we’d assumed AI Max was crucial for tapping into Google’s AI search surfaces.

    Ginny highlighted that campaigns with broad match keywords are still eligible for AI Overviews and AI Mode. Even so, AI Max does broaden possibilities by treating phrase and exact match keywords with broad match behavior and enabling keywordless matching.

    This means there are still multiple avenues available for us to access AI Search inventory.

    AI Search Reporting is Still on Hold

    Many of us were eagerly hoping for detailed reporting on AI-powered search results. However, Ginny confirmed that current ads in AI Overviews and AI Mode are reported like other top-of-page ads, with no distinct breakdown. Google’s still figuring out what these reports should eventually look like.

    This leaves us with limited insights into how much AI-driven traffic and performance we’re actually seeing.

    Google’s AI Brief: A New Layer of Control

    A significant part of the discussion circled around AI Brief, set to become the control layer for AI Max campaigns. Advertisers like me will soon be able to provide specific guidance such as “never mention prices” or define target audiences, message themes, and search intents to prioritize.

    The rollout will start with English Search campaigns and eventually spread to Performance Max and Shopping campaigns.

    For those of us worried about automation reducing our control, AI Brief offers a promising solution.

    The Core of Effective Advertising: First-party Data

    If there’s anything I walked away with, it’s the emphasis on data quality, particularly first-party data. Google’s focus is what they call “Data Strength,” and tools like Enhanced Conversions and Google Tag Gateway are pivotal.

    It’s clear: better data enhances AI performance and outcomes.

    Exploring New Metrics: Qualified Future Conversions

    Another fascinating development is Qualified Future Conversions (QFC). This metric estimates potential conversions occurring within 180 days post-ad interaction. It’s especially useful if you’re in B2B or lead generation sectors with lengthy sales cycles.

    Currently, it’s in testing with select advertisers, and I’m keen to see it roll out further later this year.

    Key Areas of Excitement at Google

    When asked about her personal highlights from GML, Ginny shared three areas: the new ad formats for AI Search, measurement innovations like QFC, and YouTube Creator Partnerships.

    This truly illustrates where Google is investing: AI discovery, advanced measurement, and creator-driven advertising.

    Putting It All Together

    This Q&A has definitely filled in some gaps left by the GML presentations. I’ve realized that broad match terms still provide a pathway to AI Search, AI-specific reporting is evolving, and Google’s vision continues to be centered on automation, powered by first-party data.

    Most importantly, it’s about balancing automation with new controls like AI Brief to shape Google’s AI systems to our advantage.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • Master AI Search Visibility: Track Influence Beyond Clicks

    Master AI Search Visibility: Track Influence Beyond Clicks

    The journey from discovery to decision is becoming increasingly obscure. I’ve discovered how to merge traditional attribution methods with new, subtle signals of influence.

    Most traditional attribution models were designed for a world where clicks were king. Someone would search for something, click on a result, visit a page, and eventually convert. Simple, right?

    Analytics platforms used to connect these actions seamlessly, painting a fairly accurate picture of success. While not perfect, at least the process was visible. Now, AI-generated search experiences have made this path much harder to trace.

    Imagine a scenario where a prospective buyer consults ChatGPT about the best project management software or leans on Google’s AI Overview for cybersecurity advice before compiling a list of potential vendors. My company might make it into those discussions without a single click to show for it. This discrepancy between influence and traffic is precisely why I need to rethink attribution.

    Search trends have been gravitating towards zero-click experiences for years now. Features like snippets, knowledge panels, and local packs have effectively reduced click-through rates by providing answers directly in the SERP.

    Generative search takes this even further by compressing what used to be a multi-click research journey into one pivotal interaction. Users can now compare vendors, appraise recommendations, and gather data without ever leaving the SERP.

    For brands, this translates to lost visibility in certain parts of the buyer journey. But it also opens up new avenues for influencing decisions before a website visit even takes place.

    Dig deeper: What 4 AI search experiments reveal about attribution and buying decisions

    Even though we’ve traditionally relied on website visits as the primary indicator that marketing has made an impact, AI is changing the game by disconnecting discovery from measurable traffic.

    A prospect might come across my brand several times through AI-generated answers before ever arriving on my site. By the trip they make to my site, their journey can look deceptively simple in analytics: Direct visit, branded search, conversion.

    Those early interactions that introduced my brand or influenced a buying decision can remain invisible in reporting.

    As more initial discovery and evaluation happens within AI frameworks, traditional attribution captures less of the decision-making landscape. While it still records visits, much of what occurs before that remains unseen.

    These harder-to-measure interactions are still crucial, creating fresh chances to influence how buyers discover, evaluate, and compare choices.

    ```json
{
  "alt": "The CapmatchOne logo with a gradient circle and bold text.",
  "caption": "Discover innovation with the CapmatchOne logo, featuring sleek typography and a modern gradient circle.",
  "description": "The CapmatchOne logo features bold, modern typography coupled with a gradient circle, symbolizing connection and innovation. The sleek design conveys a sense of progress and creativity. This image can be used for branding or promotional purposes, appealing to audiences interested in innovative solutions and forward-thinking designs."
}
```

    A potential buyer might first hear about my company through one of these AI channels, then go on to use AI to weigh options, explore alternatives, and make a shortlist—all before visiting my site. During this process, they might encounter my brand through various touches such as recommendations, comparisons, citations, and AI-generated responses that foster familiarity and build credibility.

    These interactions, despite not generating a click, can play a critical role in shaping buyer decisions and determining which brands make it to the final evaluation stage.

    Dig deeper: Why AI visibility starts before search and ends with citations


    While traditional attribution is still valuable, it now provides a less comprehensive description of how decisions are made. As AI becomes a bigger part of how buyers research and scrutinize options, a broader view of influence is essential. This involves going beyond the conversion path to incorporate signals that outline how awareness and consideration develop over time. Here’s where I begin.

    1. Assisted conversions: AI-generated recommendations frequently shape decisions well before entering a measurable funnel. Assisted conversion reports can highlight which channels influence conversions, even if they’re not the final touchpoint.

    2. Branded search growth: An observable rise in branded search activities can indicate that AI visibility is growing brand awareness. More searches for my company following AI-generated mentions are a promising sign.

    3. Direct traffic trends: While direct traffic shouldn’t solely represent AI’s influence, unexplained increases can be telling. They may suggest that people are learning about my business from AI sources before returning directly or via branded searches later.

    4. Brand visibility within AI systems: Observing how often my brand appears in AI prompts and recommendations provides valuable insight. It reflects whether AI frameworks consider my brand a credible option within a given category.

    The ultimate goal is to integrate traditional attribution data with these new visibility and influence signals to create a fuller understanding of decision-making dynamics.

    Dig deeper: The micro-macro shift: How to measure AI visibility now that precision is gone

    The takeaway here is to build a more comprehensive view of influence. My understanding of market influence starts with the realization that the consumer journey extends well beyond visible interactions and analytics.

    As AI continues to grow in prominence for discovery and evaluation, adapting strategies to account for this broader scope of influence will be crucial for staying competitive.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • Microsoft Unveils Web IQ: Revolutionizing AI-Agent Searches

    Microsoft Unveils Web IQ: Revolutionizing AI-Agent Searches

    I’m excited to share that Microsoft has launched a groundbreaking search service specifically designed for AI-agents, as agents have unique search requirements compared to humans.

    I’ve learned that Microsoft’s latest innovation, Web IQ, is here to bridge AI systems with real-time intelligence online. As a suite of AI-native grounding APIs, Web IQ sources fresh data, be it web pages, news, images, or videos, as announced by Microsoft here.

    What is Web IQ? Web IQ is all about connecting AI systems to real-world updates, leveraging Bing’s index for superior understanding. I find it fascinating how it uses the same infrastructure as Microsoft Copilot and other leading LLMs, like ChatGPT.

    However, I discovered that Web IQ’s APIs are newly developed for efficiency and relevance, crucial for serving Bing, Copilot, and ChatGPT queries rapidly.

    For AI-Agents, Not Humans. Web IQ tailors search results specifically for AI-agents. Unlike human-oriented Bing Search, ranking isn’t a priority here, as agents need swift information extraction, as stated by Jordi Ribas, President of Search & AI at Microsoft.

    Unlike us, AI-agents don’t just issue a single query; they delve deeper and continuously expand their search. This paradigm shift meant re-architecting search from indexing to orchestration, aligning it with AI needs, as per Microsoft’s insights.

    Given the frequency of searches AI-agents perform, Microsoft designed Web IQ to operate efficiently, minimizing token usage to deliver better and faster results. It’s currently 2.5 times faster than its nearest competitor.

    Access and Availability. At present, Web IQ supports Microsoft Copilot, OpenAI’s ChatGPT, and other large LLM platforms. As Microsoft scales this technology, I expect wider access to follow.

    If you want to express interest in Web IQ, Microsoft encourages you to visit this page.

    Why this Matters. As we witness the web transforming to accommodate agentic technologies, keeping an eye on these developments is vital. Websites, including mine, must evolve alongside these AI advancements.

    AI-agents aren’t just a trend; they’re part of the web’s next evolution. I’m preparing to embrace this change, and I suggest you do too.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • Google’s New AI-Powered Healthcare Ads: What You Need to Know

    Google’s New AI-Powered Healthcare Ads: What You Need to Know

    Hi there! Today, I’m thrilled to share some intriguing news about Google and its latest venture into AI-powered search advertising. Google has kickstarted testing for healthcare ads in AI Mode. This exciting development gives us a glimpse into the future of advertising within AI-based search environments.

    The scope of this test is currently narrowed to healthcare advertisers in the United States and focuses only on English-language queries in AI Mode, as confirmed by Google Ads Liaison Ginny Marvin.

    Amidst swirling industry rumors, it has now been officially confirmed that healthcare ads have indeed begun appearing in AI-generated search results.

    What Google is saying. Addressing inquiries on LinkedIn, Marvin highlighted that Google has “begun a small test of ads in AI Mode specifically for the healthcare sector.” It’s an intriguing move, isn’t it?

    She mentioned that a variety of campaign types can participate, including:

    • Performance Max (PMax)
    • AI Max with search term matching
    • Shopping campaigns
    • Broad match campaigns

    These campaign types can also show ads within AI Overviews.

    ```json
{
  "alt": "LinkedIn conversation about Google testing healthcare ads in AI Mode.",
  "caption": "Exploring the future of ads: A LinkedIn dialogue on Google's AI Mode for healthcare advertising.",
  "description": "A LinkedIn exchange between Ben Goldman and Ginny Marvin discussing Google's test of AI Mode ads in the healthcare sector. Ben Goldman inquires about the testing scope, seeking clarity on initial restrictions like creatives without pinned assets or text disclaimers. Ginny Marvin confirms the test's scope, highlighting that it's limited to the US for English queries and involves PMax and AI Max. The conversation reveals insights into Google's advertising innovations."
}
```

    Why we care. As healthcare stands as one of Google’s stringently regulated advertising sectors, this test is crucial for understanding how Google might monetize AI-driven search results. If the test expands, healthcare marketers could gain a new platform for visibility, and advertisers in similarly regulated industries might get a sneak peek of future ad appearances in Google’s AI-generated search.

    The fine print. This initial testing phase comes with some creative boundaries. Marvin noted that ads devoid of pinned assets or text disclaimers are presently the only eligible healthcare advertisements.

    What to watch. It’s just the beginning, and we’re curious to see if Google will broaden the eligibility for more healthcare ads, introduce other ad formats, or extend into other regulated fields.

    Such developments could provide early clues about Google’s strategy to harmonize monetization and user trust as AI Mode starts to play a more significant role in the search experience.

    First spotted. Interestingly, it was Senior Strategist Ben Goldman who first noticed this test, which he shared in response to her GML 2026 summaries on LinkedIn.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • Unveiling Intelligent AI Search: The Future of Content Visibility

    Unveiling Intelligent AI Search: The Future of Content Visibility

    Have you ever wondered how AI search platforms have evolved from simple Retrieval-Augmented Generation (RAG) to sophisticated agentic systems? These days, AI search has advanced beyond mere RAG, transforming into something far more complex and dynamic. In this article, I’ll guide you through how today’s advanced AI retrieval systems determine if your content is showcased or left in the shadows.

    About two and a half years ago, I penned an article for Search Engine Land on how RAG represents the future of search. It wasn’t just a reactionary measure from Google in response to ChatGPT, but rather an architecture in development since the REALM paper in August 2020. Observing developments since then, everything has aligned with what I speculated.

    ```json
{
  "alt": "Illustration showing process breakdown: query, bad first pull, data request via vector search, distinction between fake and real.",
  "caption": "Exploring why naive RAG models fail: A journey through confusing queries, poor data pulls, and the challenge of distinguishing fake from real.",
  "description": "This image illustrates the breakdown of a naive RAG (Retrieval-Augmented Generation) process. It features four panels: the first shows a query with connections, the second highlights a 'bad first pull' within an orange target, the third depicts a 'data request' and 'vector search', and the fourth illustrates a spiral symbolizing the distinction between 'fake' and 'real' data. The image conveys complexities in data retrieval and processing, serving as a cautionary tale for content marketers."
}
```

    The RAG pipeline of the past, which I outlined as a query transforming to an answer with citations, is already outdated. Major AI search platforms like Google AI Mode and ChatGPT Search have transitioned to a more complex architecture. They now possess planning capabilities, tool-routing options, and iterative retrieval methods that continuously refine results until they reach a suitable conclusion. The one-retrieval-to-answer model is defunct.

    ```json
{
  "alt": "Illustration of a user query process showing a planner and sub-queries branching from a main query.",
  "caption": "Visual representation of an agentic RAG process: a user query flows to a planner, branching into structured sub-queries.",
  "description": "This illustration depicts a user at a computer initiating a 'User Query' that connects to a 'Planner'. The planner organizes multiple 'Sub-queries', represented as branches with arrows pointing to folders. This visual explains the concept of agentic RAG in handling complex queries. Keywords include user query, planner, sub-query, and agentic RAG."
}
```

    This sophisticated approach is what we now refer to as agentic RAG, a framework that’s become the industry standard. If your content strategy still relies on single-shot retrieval, you’re optimizing for a non-existent system. What’s more, in agentic RAG, you can’t witness the gatekeeping process—only the final outcome shows if your content made it.

    ```json
{
  "alt": "Comparison of Classic RAG and Agentic RAG processes.",
  "caption": "Explore the dynamic evolution from Classic RAG to Agentic RAG, highlighting enhanced retrieval and synthesis for more effective answers.",
  "description": "This image contrasts Classic RAG and Agentic RAG methodologies. The Classic RAG process involves a query leading to a smart search connected to a Large Language Model (LLM) and a private knowledge base, producing an answer. In contrast, Agentic RAG uses retrieval tools, a critic, and a synthesizer, allowing for more complex planning and routing before delivering an answer. This diagram emphasizes the improved capabilities in modern RAG approaches."
}
```

    By the time you finish reading, you’ll have a functional understanding of agentic RAG, the patent evidence showing its application by companies like Google, insights into what each major platform is doing, and concrete tactics to enhance your content strategy. You’ll also gain my important takeaway of the year: the future hinges on model distillation.

    ```json
{
  "alt": "Diagram titled 'The Agentic RAG Reference Architecture', showing vector database, live web fetch, router, code interpreter, and structured API.",
  "caption": "Explore the Agentic RAG Reference Architecture—a streamlined flow from vector database to structured API, highlighting efficient data handling.",
  "description": "This diagram, titled 'The Agentic RAG Reference Architecture', outlines a system flow from a vector database, through live web fetch, a central router, code interpreter, and finally to a structured API. The connectivity is visualized with bold yellow lines, and each stage is marked with corresponding icons and labels. Ideal for visualizing advanced data architecture, this image is designed for tech and marketing professionals seeking streamlined solutions."
}
```

    The October 2023 perspective is still relevant. Passage-level retrieval remains essential to relevance, and knowledge graphs work in tandem with LLMs. Search systems aim to lower what are known as Delphic costs, minimizing the effort users expend to find answers. Google’s guiding principle has always seen traffic as a means rather than an end. This aspect of my argument needs no change.

    ```json
{
  "alt": "Illustration of Critic/Reflection Module transforming biased and old content into fresh, diverse output through a synthesizer.",
  "caption": "Transform outdated content using the Critic/Reflection Module, turning biased and stale ideas into fresh, diverse perspectives.",
  "description": "This illustration depicts the Critic/Reflection Module process, where salesy or biased and stale or old documents are filtered into a funnel. The process refines these inputs, represented by a thumbs-up circle, into diverse and fresh content. The final output is synthesized, illustrated as a sparkling document. Keywords: Critic/Reflection Module, content transformation, synthesizer, diversity in content."
}
```

    What has evolved is the structure of the retrieval pipeline. Back in 2023, RAG was straightforward and linear. A query was encoded, top passages were retrieved, and an answer was generated. If your content was within the top set of results, you had visibility; if not, you were invisible. This framework served its purpose at the time.

    ```json
{
  "alt": "Diagram illustrating pairwise ranking of content fragments with LLM judge and synthesizer.",
  "caption": "Explore the rigorous process of content evaluation, where a powerful LLM judge analyzes pairwise content fragments, selecting the superior option for synthesis.",
  "description": "This image depicts a flowchart explaining the pairwise ranking of content fragments. Two documents, A and B, are evaluated by an LLM 'Judge', which selects the preferred document chunk, marked as Chunk A, based on a checkmark. This superior chunk is then processed by a 'Synthesizer'. The design emphasizes scrutiny in content generation, with the tagline 'Your content must survive pairwise scrutiny'. Keywords: content ranking, LLM, synthesizer, pairwise evaluation."
}
```

    Today’s pipelines boast abilities absent from linear models: planning, tool usage, multi-hop iteration, and reflection. Rather than being a single occurrence, retrieval now involves up to twenty sub-retrievals orchestrated by a central agent, which refines its foundation of evidence continuously before finalizing an answer.

    ```json
{
  "alt": "Diagram of Canonical Bridge with entities A and B connected by a content bridge.",
  "caption": "Illustration of a Canonical Bridge linking entities A and B, symbolizing a strategic content marketing approach.",
  "description": "This image illustrates a conceptual framework called the Canonical Bridge, where Entity A and Entity B are linked by a content bridge. A blue icon with a robot symbol highlights a key aspect of content marketing strategy. The diagram visually represents the transition and connection between two entities, emphasizing the role of strategic content in bridging gaps. Keywords: Canonical Bridge, content marketing, entities, strategic connection."
}
```

    My earlier writing hinted at these upgrades without naming them precisely.

    ```json
{
  "alt": "Diagram comparing a long-form document to a structured API tool with a router in between.",
  "caption": "Navigating the choice between comprehensive guides and efficient API tools: which path will your strategy take?",
  "description": "This image illustrates a comparison between using a detailed, long-form document (ultimate guide with 2500 words) and a structured API tool. The illustration shows a 'router' that routes between 'skip' and 'call' options, depicting decision-making in content strategy. Ideal for visualizing choices in content marketing, the diagram uses icons and text for clarity."
}
```

    The word “agentic” is used liberally, but its structural definition is specific. Understanding agentic RAG requires grasping four properties each system must embody to wear the label.

    ```json
{
  "alt": "Illustration showing data transfer between a Production AI unit and a Distilled Local Agent.",
  "caption": "Visualizing the seamless data flow between advanced Production AI and its streamlined Distilled Local Agent counterpart.",
  "description": "This illustration depicts a technological concept with two main structures: a large gray 'Production AI' unit on the left and a smaller transparent 'Distilled Local Agent' on the right. Colored lines between the two boxes symbolize data transfer, suggesting interaction and processing. The design highlights AI and automation, emphasizing efficiency and innovation in data handling."
}
```

    1. Planning involves restructuring the user query into a research plan, breaking it down into sub-queries, pre-selecting tools, and strategizing retrieval sequences. The system doesn’t just respond; it plans each step with precision.

    ```json
{
  "alt": "Dashboard displaying new KPIs with circular graphs showing sub-query coverage at 87%, reflection survival rate at 68%, pairwise win rate at 72%, and tool-call inclusion at 0.41.",
  "caption": "Explore key performance insights with this dynamic dashboard, showcasing metrics like sub-query coverage at 87% and a 68% reflection survival rate. Dive into data-driven success!",
  "description": "This image features a detailed KPI dashboard highlighting four metrics: sub-query coverage, reflection survival rate, pairwise win rate, and tool-call inclusion. The sub-query coverage is represented as a circular graph at 87%, with 391 queries covered out of 450. The reflection survival rate graph, labeled 'High Survival', indicates 68% over seven days. The pairwise win rate is 72%, comparing Model A (72) and Model B (27). Tool-call inclusion shows a rate of 0.41 with 112 successful out of 273 attempts. This dashboard is designed for content marketing insights."
}
```

    2. Tool usage extends beyond basic retrieval to include inquiries through APIs, code execution, live web browsing, and more. The agent selects the best method for each task, weaving these tools into cohesive outputs.

    ```json
{
  "alt": "Code snippet showing commands for cloning a GitHub repository and setting up a Python environment.",
  "caption": "Quickly set up your development environment with these concise Git and Python commands!",
  "description": "This image displays a code snippet for cloning a GitHub repository 'agentic-rag-distillation'. It includes commands to navigate into the directory, install dependencies from 'requirements.txt', pull resources using 'ollama', and copy an environment example file. The final line provides a reminder to fill in 'SERPAPI_KEY' and 'BRAND_DOMAIN'. This is ideal for developers setting up a new project environment."
}
```

    3. Iteration or multi-hop retrieval is where the agent refines its findings by visiting the source multiple times, continually improving the evidence base.

    ```json
{
  "alt": "Code snippet showing a Python command to run an audit with brand domain and trace output options.",
  "caption": "Running an audit has never been easier with this Python command. Customize your query, brand domain, and trace output to streamline your tasks.",
  "description": "This image features a Python command used to perform an audit. It includes options to input a specific query, a brand domain URL, and specifies the trace output file path. Useful for developers looking to automate audits with customizable inputs, this snippet demonstrates command-line flexibility and efficiency in running tasks. Keywords: Python, audit, command-line, automation."
}
```

    4. Reflection involves the agent critiquing its own output, determining its sufficiency and quality, and retrieving more information if needed to resolve discrepancies or improve source diversity.

    ```json
{
  "alt": "Screenshot of an AI-driven query resolution process displaying data retrieval and evaluation results.",
  "caption": "Exploring AI-driven query fan-out: A detailed look into how complex search queries are broken down and evaluated for optimal results.",
  "description": "This image showcases a comprehensive overview of the AI-driven query fan-out process, demonstrating how complex queries are broken into sub-queries for efficient data retrieval. The screenshot includes retrieval funnel statistics, pairwise decisions, and critique notes, reflecting the intricate mechanisms used to enhance the accuracy and relevance of search results. Key elements include website rankings, query routing reasons, and citations, providing a detailed framework for understanding AI query operations."
}
```

    These are the qualities that set agentic RAG apart and what make it the new default for AI search platforms.

    ```json
{
  "alt": "Python command with options for trace directory and brand domain in code snippet.",
  "caption": "A Python command ready to execute a view program with specified trace directory and brand domain options.",
  "description": "This image features a code snippet formatted in XML style, showcasing a Python command to run a module named 'examples.view_program' with options for setting a trace directory to 'traces/' and a brand domain as 'yourbrand.com'. The command includes newline escapes for readability. The code snippet is enclosed in XMP tags, indicating a block of computer code."
}
```

    Drawing a contrast between the classic RAG and agentic RAG, imagine the former as a direct process and the latter as a comprehensive loop where steps can be revisited until the solution is optimal. This is what my content needs to withstand.

    ```json
{
  "alt": "Screenshot showing metrics and query processing output for a relevance engineering task.",
  "caption": "A glimpse into the evaluation metrics and query processing steps in relevance engineering using a brand-specific retrieval task.",
  "description": "This image captures a terminal screenshot displaying metrics and outputs from a relevance engineering task. Metrics such as sub-query coverage, retrieval-to-citation ratio, and reflection survival are presented. It includes a stage-failure rate table with failure stage data, and a per-query funnel showing progression or failure across different query processing stages. Keywords like 'relevance engineering', 'query processing', and 'retrieval metrics' are explored in the context of brand processing for ipullrank.com."
}
```

    The six shifts required for effective content engineering in the realm of agentic RAG are clear. I need to optimize for a spectrum of sub-retrievals, present well-structured and cohesive passages, leverage bridge entities, offer tool-callable content, and ensure freshness within my content.

    ```json
{
  "alt": "Code snippet illustrating a Python command for comparing local and production files.",
  "caption": "Exploring file comparisons: This Python command snippet demonstrates how to compare local traces with production files using YAML configurations.",
  "description": "The image displays a code snippet within an 'xmp' tag, showcasing a Python command. This command compares local JSON trace files against production YAML files. It's a useful tool for developers to ensure consistency and correctness across different environments. Keywords: Python command, file comparison, JSON, YAML, script."
}
```

    The path forward involves navigating measurement’s increasingly complex landscape with the aid of model distillation. By understanding the full lifecycle from internal query generation to external execution, I can effectively target content positioning and citation strategy.

    Engaging with this agentic environment demands observation, adjustment, and perpetual calibration. The choice is simple: evolve to survive and thrive or remain static and risk obscurity.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • 5 Essential AI Search Insights Every Marketer Should Know

    5 Essential AI Search Insights Every Marketer Should Know

    I recently dove into five fascinating studies that are truly changing the way we as marketing leaders approach AI Search. These insights are not only reshaping our strategies but also pushing us to think beyond traditional SEO methods.

    Each study offers a unique perspective on how AI can enhance search capabilities, enabling us to connect with our audience more effectively. It’s exciting to explore how these powerful tools can transform our marketing efforts.

    By understanding these groundbreaking research findings, I feel more empowered to make informed decisions that align with the evolving digital landscape. The integration of AI is inevitable, and embracing it will help us stay ahead of the curve.

    The challenge is clear: we must integrate these insights seamlessly into our marketing strategies. Doing so will enhance our ability to deliver personalized and impactful content to our audience, fostering deeper engagement and driving success.

    I’m eager to see how these studies will continue to influence and define marketing practices, leading to more innovative approaches and ultimately, better results for our brands.


    Inspired by this post on Try Profound Blog.


    crushpress.ai community screenshot
  • Unlocking AI Visibility: The Key Role of Brand Depth

    Unlocking AI Visibility: The Key Role of Brand Depth

    Have you ever wondered why some brands consistently show up in AI recommendations, while others don’t? I’ve discovered that building deep and consistent brand presence is the real game changer.

    I’ve come to realize that simply getting cited isn’t enough. It’s the brands with a strong semantic footprint the AI systems love to retrieve and recommend.

    For me, generative engine optimization (GEO) is like playing two games at once: creating both long-lasting brand influence within AI systems and crafting content that navigates modern data retrieval pipelines effortlessly.

    During my deep dive into AI recommendations, I learned that brand depth significantly boosts your chances in both retrieval and synthesis processes.

    Playing Two Games: The GEO Challenge

    Every layer I explored influenced visibility differently.

    Game 1: Building Parametric Weight

    Brands are like coordinates in a language model’s embedding space, shaped by the density and consistency of signals. I’ve found that building this weight takes time, growing steadily over months, even years.

    ```json
{
  "alt": "Comparison of low and high entity depth for Brand X and Black Honey by Clinique.",
  "caption": "Exploring Brand X with low entity depth versus Clinique's Black Honey, featuring high entity connections including Liv Tyler and TikTok.",
  "description": "The image illustrates a comparison between Brand X with low entity depth and Black Honey by Clinique with high entity depth. Brand X shows limited connections with question marks, signifying weak market pull. In contrast, Black Honey is linked with specific entities such as MLBB, Liv Tyler, TikTok, and the year 1971, indicating strong market influence. This visual emphasizes the significance of brand associations in consumer appeal."
}
```

    A brand with inconsistent messaging, as I’ve seen, ends up with a fuzzy vector, which hampers recall and confidence during AI retrieval.

    Through my experiences, it’s clear that ignoring the foundational elements of a brand in favor of short-term citation strategies leads to missed opportunities in AI systems’ recognition.

    Game 2: Survival of Retrieval

    For me, the true test comes when systems like Google AI Mode or ChatGPT Search launch their retrieval pipelines. Will my content make it through? About 85% of brand mentions in AI systems stem from external domains, which says a lot about where I need to focus my efforts.

    Different AI search systems have their unique methods, from Perplexity’s citation embedding to Google’s query fan-out, and each presents its own set of challenges and opportunities.

    Citations: Just the Surface

    In my findings, citations only signal output presence, not the underlying retrieval and synthesis processes. Focusing solely on citations can be misleading. It’s important to delve deeper into the factors that lead to citation in the first place.

    ```json
{
  "alt": "The CapmatchOne logo with a gradient circle and bold text.",
  "caption": "Discover innovation with the CapmatchOne logo, featuring sleek typography and a modern gradient circle.",
  "description": "The CapmatchOne logo features bold, modern typography coupled with a gradient circle, symbolizing connection and innovation. The sleek design conveys a sense of progress and creativity. This image can be used for branding or promotional purposes, appealing to audiences interested in innovative solutions and forward-thinking designs."
}
```

    Brand Depth: The Familiar Route for AI and Humans

    As I looked into it, I realized that human brains and LLMs share a common strategy: defaulting to the familiar through dense information frameworks.

    Predictive processing theory helped me understand why both prioritize densely established information, highlighting the similarities between human decision-making and AI functions.

    Getting Technical with Brand Depth

    Diving into the technical aspects, I learned that Google and AI models focus on entity salience, coherence, and relational density to determine a brand’s visibility and reliability.

    Entity Salience

    I discovered that high entity salience increases the likelihood of being cited and recognized in AI systems.

    Low salience restricts visibility to exact branded queries, whereas high salience ensures my brand surfaces even when just the topic is discussed.

    ```json
{
  "alt": "Four thumbnails of short makeup tutorial videos, each showing a different woman applying Clinique products.",
  "caption": "Explore the latest Clinique makeup trends with short tutorials showcasing different products and application techniques. Perfect for beauty enthusiasts!",
  "description": "This image displays four thumbnails of short videos, each featuring a makeup tutorial. The videos highlight different women using Clinique cosmetics, such as lipsticks and skincare products, with varying durations between 10 to 91 seconds. These tutorials are sourced from popular social media platforms like Instagram and TikTok. The image provides a glimpse into contemporary makeup practices and the use of specific products for achieving different looks. Keywords: Clinique, makeup tutorial, short videos, beauty, cosmetics, Instagram, TikTok."
}
```

    Entity Coherence

    I’ve realized the importance of maintaining a consistent brand identity to avoid low confidence representations in AI models, which otherwise leads to brand drift over time.

    Inter-entity Relationship Density

    Building strong connections with authoritative entities enhances the chances of my brand being retrieved and recognized during AI reasoning processes.

    The RAG Layer: Where Site Quality Shines

    I’ve learned from Mark Williams-Cook that a site’s quality score can determine its eligibility for retrieval, emphasizing the need for strong brand infrastructure for consistent visibility.

    Why AI Systems Highlight Clinique’s Black Honey

    Clinique’s “Black Honey” lipstick is a fantastic case. Its impressive entity depth frequently registers it in AI responses. I aspire for such widespread recognition for my endeavors.

    From its cultural anchors to competitive benchmarking, the layers of meaning around “Black Honey” continually rack up its mentions and trustworthiness in AI systems.

    Crafting Content for AI Retrieval Success

    In my approach, focusing on rich, unique content is crucial. High-quality content naturally finds its way through the retrieval funnel, while generic content falls by the wayside.

    By crafting detailed, data-rich narratives, I ensure that my work stands out as essential, enhancing chances of being cited and referenced by AI tools.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • Unveiling the SEO-GEO Divide: AI Traffic vs. Organic Traffic Secrets

    Unveiling the SEO-GEO Divide: AI Traffic vs. Organic Traffic Secrets

    The SEO-GEO gap- How AI search traffic differs from organic traffic

    Looking at data from 10 websites, I discovered why original research, innovative tools, and answer-focused content often outperform generic educational articles in the GEO realm.

    Some marketers believe GEO might replace SEO, while others say robust SEO is enough for AI visibility. So, I decided to dig into both perspectives by examining LLM referral traffic and organic traffic across 10 different sites.

    Here’s what I found out about how AI search leans towards specific content patterns that differ from traditional organic search.

    3 Key Findings from the Dataset

    1. Traditional SEO Content Strategies Fall Short for GEO

    I noticed blog content themes were a strong predictor of LLM traffic. Educational “comprehensive” guides often underperformed compared to shorter posts with unique data.

    Trends and analysis posts were cited by LLMs 78% of the time. Posts featuring unique data held a significant lead in the citation pool, while educational how-to content lagged behind at a mere 12%.

    It became clear that producing content rich in data and measurements significantly boosts your chances of entering the LLM citation pool. On the other hand, generic educational content might not make the cut.

    2. Organic Success Doesn’t Ensure LLM Traffic

    In my analysis, the top 10 organic pages captured over half the organic sessions but only 29% of LLM sessions.

    Your most successful organic content may not necessarily perform well with LLM traffic. Among the top 100 organic pages, nearly half didn’t receive any LLM traffic at all!

    Although there’s some correlation between organic performance and LLM traffic, the two aren’t equivalent.

    3. Service/Product Pages Excel in LLM Traffic

    While articles and blogs brought in most LLM referrals by session count, service and product pages outperformed others when LLM sessions are considered per 1,000 organic sessions, making them significant performers.

    Page typeLLM sessions per 1,000 organic
    Service/product29.4
    Article/content23.4
    FAQ/support14.0
    Tool/demo9.8
    Homepage5.6

    Turning my attention to practical insights, it was evident that crafting authoritative content that offers specific answers can significantly enhance LLM traffic. Integrating interactive tools emerged as another powerful approach. When LLMs recommend tools, they drive targeted traffic effectively.

    The Methodology Behind My Case Study

    I analyzed GA4 data from 10 diverse websites, covering 150,000 indexed pages in March 2026 to gather these findings.

    • The domains, handpicked for their varied industries and consistent SEO performance, ranged across healthcare, technology, retail, and more, ensuring a balanced view.
    • I meticulously isolated LLM-referral traffic using GA4 channel groupings and segmenting referrer paths, focusing on sessions from major AI platforms like ChatGPT.
    • Content type categorization helped me compare LLM citations, while I used per-page averages from GA4 for engagement time analysis.

    It’s worth mentioning that LLM bot crawls aren’t captured by GA4, as they make server-level requests before client-side JavaScript loads. Thus, the organic session data reflects only human visitors.

    What LLM Traffic Patterns Reveal About Engagement

    ```json
{
  "alt": "The CapmatchOne logo with a gradient circle and bold text.",
  "caption": "Discover innovation with the CapmatchOne logo, featuring sleek typography and a modern gradient circle.",
  "description": "The CapmatchOne logo features bold, modern typography coupled with a gradient circle, symbolizing connection and innovation. The sleek design conveys a sense of progress and creativity. This image can be used for branding or promotional purposes, appealing to audiences interested in innovative solutions and forward-thinking designs."
}
```

    LLM Referral Behavior vs. Organic Traffic

    Analyzing engagement time across traffic types revealed averages were similar—yet disparities emerged across different page types.

    Page typeOrganic avg. timeLLM avg. time
    Tool/demo101 seconds146 seconds
    Homepage36 seconds82 seconds
    Service/product69 seconds63 seconds
    Article/content56 seconds40 seconds

    Tools and homepage content saw heightened engagement from LLM users, suggesting they look for actionable insights rather than merely seeking information.

    Recognizing the Potential of Interactive Tools with LLM Traffic

    Interactive tools received the highest per-page LLM citations, and these tools were prominently featured by LLMs in response to relevant user queries.

    Emergence of LLM-only Traffic

    Interestingly, some LLM-receiving pages recorded no organic clicks, which could signify unique discovery mechanisms. This study showed engagement quality on these pages was notably high, driven by LLM-directed users ready to engage.

    GEO Tactics Supported by Data

    Answer Questions LLMs Can’t Address Themselves

    It was evident that generic educational content is often redundant for LLMs. Content differentiation comes from original research and proprietary insights.

    Investing in research and verifiable data can significantly enhance your content’s GEO impact.

    Implement Answer Capsules

    Research shows answer capsules, concise responses placed prominently, are strongly favored by LLMs for citation.

    By providing direct answers early, the pages excelled in LLM traffic.

    Maximize Named Interactive Tools

    If your site includes calculators or assessments, highlight them for GEO success. Ensure they are easily found and provide valuable, targeted insights.

    Separate Tracking for Organic and LLM Pages

    Recognizing that organic and LLM hits don’t always align, thoughtful mapping based on AI queries can reveal high-quality LLM traffic opportunities.

    Pages that solely receive LLM attention can still hold value, as users arrive prepared for deeper engagement, driven by AI direction.

    Same Strategies, Different Tactics in GEO and SEO

    This analysis highlighted that while GEO coexists with SEO, it demands distinct page tactics. As zero-click searches grow, understanding and leveraging these nuances becomes crucial.

    By constructing content that answers specific questions with original data and strategic uses of GEO tactics, you can optimize for both systems. Keep in mind, mastering one does not automatically ensure success in the other.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot
  • How Google’s AI Evolution Will Reshape Search and the Web

    How Google’s AI Evolution Will Reshape Search and the Web

    I recently followed an intriguing conversation with Google’s CEO, Sundar Pichai, where he explored the transformative journey that awaits Google’s AI, Search, and digital tools. The path forward envisions these elements coalescing into a unified powerhouse capable of executing tasks seamlessly.

    In a detailed exchange with Nilay Patel from The Verge, Pichai addressed concerns about an evolving Search landscape. He firmly reiterated Google’s commitment to connecting users with the open web, assuaging publisher concerns about potential traffic declines.

    Pichai assured, “Through it all, we are very committed to both meeting user expectations and also connecting them to what’s out on the web.” Yet, it’s clear why some fears persist as Google steers towards an AI-driven future where Search evolves to include conversational agents and task-oriented tools, reducing the need for traditional clicks.

    Why we care. It’s important to recognize the emerging landscape, one where Google’s Search, Gemini, and agent technologies blend into a singular AI layer. This shift points toward a revamped approach to discovering information, creating content, and handling tasks.

    Agents are the future. These AI agents are poised to drive the next evolution on the web. According to Pichai, “I look at agents, and that is the next evolution of the web. I think it will evolve the web pretty profoundly.”

    In the background, Google’s efforts in developing agentic tools across Search, Gemini, Spark, and Antigravity aim to bring these innovations together for a more cohesive user experience. Acknowledging this unified trajectory, Pichai envisions Google’s ecosystem as evolving into an ‘agent manager’ model.

    One product. When asked if Google’s suite of AI search and app-building tools might eventually merge into one, Pichai affirmed, “It will.” This convergence means Google agents will quietly assist users in planning and executing tasks, a vision for which Google is diligently assembling essential building blocks.

    Pichai elaborated, “We are laying a lot of the primitives of what we need for agents to work end to end, and more importantly, for AI to work.”

    Dig deeper. Explore perspectives on how Google’s Search and Gemini might converge or continue to diverge in the discussion led by Google’s Liz Reid.

    Google rejects Google Zero. In the face of concerns about Google’s evolving role in web traffic, Pichai illustrated his view of an expansive information ecosystem, far broader than Google alone.

    Addressing Condé Nast’s apprehension about declining search traffic, he highlighted the dynamism of the current landscape, where publishers adapt continually to shifts in user behaviors and new digital formats.

    “It’s exceptionally dynamic, and so it makes sense to me every publisher is adapting to this new world,” he observed.

    Google says some clicks are going away. While Pichai refrained from advising publishers on business planning, he emphasized that as technology improves, low-quality clicks naturally dwindle, alongside metrics reflecting a decline in bounce clicks.

    Google points to subscriptions. By highlighting Google’s adjustments to support subscription models, Pichai acknowledged this as a key adaptation amid evolving publisher strategies.

    “We are adapting to the fact that publishers are increasingly turning to subscription offerings, too,” he stated, promoting Google’s efforts to highlight subscribed content as preferred sources for users.

    It’s worth noting that the drive towards subscriptions was, in part, a response to diminishing reliance on search traffic.

    Search had to move faster. The decision to reorganize Google Search was a strategic move to enhance agility in the rapidly advancing AI era, positioning the platform for rapid decision-making and innovation under new leadership.

    For more insights into Sundar Pichai’s thoughts on AI, search, and the future of the web, consider listening to the full interview here.


    Inspired by this post on Search Engine Land.


    crushpress.ai community screenshot