I recently came across some eye-opening data about ChatGPT and its impact on driving traffic to publishers. The findings reveal a substantial gap between the visibility of ChatGPT links and actual clicks, which is quite astonishing.
A leaked document shows how OpenAI is monitoring user interactions, especially focusing on how frequently ChatGPT provides publisher links and the surprisingly low number of users who click on them.
By the numbers. ChatGPT does indeed feature links, yet they receive minimal engagement. For a top-performing page, here’s what the OpenAI data indicates:
610,775 total link impressions
4,238 total clicks
0.69% overall CTR
Best individual page CTR: 1.68%
Most other pages: 0.01%, 0.1%, 0%
ChatGPT metrics. This leaked file details each instance where ChatGPT displays links, providing a breakdown of user interactions:
Date range (include date partition, report month, min/max report dates)
Publisher and URL details (publisher name, base URL, host, URL rank)
Impressions and clicks across various locations:
Response
Sidebar
Citations
Search results
TL;DR
Fast navigation
CTR calculations for each display area
Total impressions and total clicks across all surfaces
Where the links appear. Surprisingly, the zones with the most visibility yield the fewest clicks. Here’s a performance breakdown by visibility zone:
Main response: Massive impressions, minimal CTR
Sidebar and citations: Reduced impressions but higher CTR (6–10%)
Search results: Negligible impressions, zero clicks
Why it matters. If you were hoping ChatGPT’s visibility could substitute for your lost Google organic search traffic, think again. Although AI-driven traffic is on the rise, it remains just a sliver of overall traffic and unlikely to match the behavior of traditional organic search traffic.
About the data. This fascinating data was shared on LinkedIn by Vincent Terrasi, CTO and co-founder of Draft & Goal, a company specializing in content production workflows.
In this report, I explore the finest B2B SaaS marketing agencies of 2025, evaluating them based on various criteria. Let’s dive into the aspects that set these agencies apart!
The companies were assessed on their notable clients, experience in the industry, the longevity of their leadership, and much more.
The most important criterion was their client base and how they’ve adapted their services to align with different market demands.
Here’s a detailed breakdown of the criteria I used to rank these agencies:
Notable Clients (20%): This is crucial to understanding an agency’s experience with B2B SaaS clients.
Year Founded (10%): Older agencies usually have the wisdom of adapting to fluctuating market conditions.
Leadership Experience Score (10%): This score reflects the expertise of the agency’s leadership in the marketing sector and their experience with B2B SaaS.
Average Reviews (20%): I normalized the agencies’ reviews from multiple third-party sources to ensure fairness in ranking.
Founder Led & Median Employee Tenure (10% each): Agencies led by their founders and with long-tenured employees signal stability and quality.
GEO Offering (10%): Agencies that offer Generative Engine Optimization (GEO) have a competitive edge in helping clients rank well in AI-generated overviews.
Media References (5%): This indicates how frequently an agency’s work is cited by authoritative media sources.
AI Visibility Score (5%): A proprietary measure of the agency’s visibility and that of its clients in AI-driven platforms like ChatGPT.
The list below showcases the top 10 B2B SaaS marketing agencies, as ranked by these factors. Additionally, I included information about their headquarters and marketing specializations.
I’ve noticed that many people labeling things as “AI SEO” are just applying traditional SEO concepts dressed up with new buzzwords.
AI SEO, however, stands apart.
When I explore how AI tools like AI Overviews, ChatGPT, and Perplexity sort and condense information, it’s clear there are strategies available to us now that simply didn’t exist in the old Google 10-blue-links era.
In this article, I’ll walk you through those unique AI SEO tactics, leveraging concrete data, not just hopeful speculation.
Feeling the drop in clicks, right? Here are some compelling facts:
Research has shown that when Google’s AI Overviews were applied, the click-through rates to top organic results fell by about 30 to 35%. In some cases, publishers reported losing 40 to 80% of their search traffic.
According to an analysis with Similarweb data, news traffic from Google declined from around 2.3 billion to under 1.7 billion visits in just a year as zero-click searches increased from 56 to 69% after AI summaries were introduced.
From a Semrush study on 10 million keywords, AI Overviews now frequently appear, especially for informational queries, changing the visibility landscape by consolidating multiple sources into a single AI-generated response.
Meanwhile, the AI market is expanding at a rate of over 30% CAGR, with projections suggesting that total AI spending will reach into the trillions by the early 2030s.
AI SEO is about optimizing not just for clicks but for factual representations that earn places within AI-generated answers.
Here are 12 exclusive tactics to thrive in this new landscape:
1. Prompt Graph Coverage
Traditional SEO treats a query as a single unit mapped to a page.
AI engines deconstruct queries into graphs of subtasks and address each. Google mentions “multi-step reasoning” for tackling complex queries at once. Academic research on AI SEO also indicates that AI functions break down queries into sub-questions, synthesizing information across sources.
AI SEO strategy: Model that graph personally.
Transform the primary query into predictable sub-questions.
Create detailed sections that fully address each subtask.
Ensure each section is self-contained and suitable for the specific micro-intent.
When writing about “best project management software,” consider prompting for:
“criteria for agencies”
“comparison vs spreadsheets”
“pricing breakdown by seat”
“implementation timeline”
Each needs its own precise, well-titled segment.
2. LLM Seeding
While traditional search engines don’t absorb all content into their algorithms, LLMs do.
AI SEO shows a preference for neutral sources like Wikipedia and governmental documents over branded marketing pages, so contributing to factual and earned sources is key. Backlinko’s findings reinforce engaging in the right content surfaces for training and retrieval.
AI SEO-only move:
Release definitions, glossaries, and FAQs publicly.
Contribute to places where models learn their foundational facts.
Sow Q&A style content in widely used forums.
This is about showing where the model will find the canonical truth, making sure it’s your content.
3. Passage-Level Retrieval Optimization
Traditional SEO generally ranks entire pages. AI engines retrieve information at a passage level.
Studies show that models cite specific highly structured passages, not entire pages.
AI SEO-only move:
Treat each heading as a standalone answer.
Include all claims, qualifiers, and evidence within one passage.
Minimize the reader’s need to traverse the page for logic.
Stand out as the model’s go-to reference for any particular question.
4. Citation-Ready Evidence Packaging
AI engines must justify their responses.
Studies indicate pages commonly cited by AI engines have structured data, semantic HTML, and explicit evidence like tables. The absence of verifiable facts increases the tendency for models to hallucinate.
AI SEO-only move:
Present data in machine-readable formats: tables, comparisons, glossaries, checklists.
Support each strong claim with solid statistics and a source.
Ensure the model can easily extract your “proof block.”
You need to be verifiable and structured for easy reuse.
5. Neutrality Engineering
Models favor neutral, non-promotional sources over overtly commercial ones.
According to research, Google’s definition of spam has widened to include content that lacks depth, especially in AI Overviews.
AI SEO-only move:
Remove promotional language from pages aimed at being cited.
Ground your narrative in facts, comparisons, and third-party validations.
Create separate layers for opinion and positioning.
Continue to sell, but ensure your main content remains neutral and evidence-based.
6. Brand-Entity Memory Alignment
While search engines focus on page-query matching, LLMs concern themselves with how well your entity is understood across the board.
Studies suggest variance in how engines perceive brands, often favoring well-recognized and consistently presented entities.
AI SEO-only move:
Clearly define your brand’s canonical facts: identity, operations, audience.
Ensure consistency across high-authority platforms.
Rectify outdated or conflicting information across channels.
Train the model to understand who you are, not just what metadata say.
7. Competitor Co-occurrence Hijacking
A significant portion of buying intent lies in comparative prompts.
AI engines synthesize answers by comparing multiple competitors. Research shows brands frequently appearing in comparative content often benefit in AI outputs.
AI SEO-only move:
Position your brand in neutral, third-party comparison content.
Craft balanced comparisons that consider multiple competitors honestly.
Encourage inclusion in “shortlist” content likely used in category training.
Traditional SEO hopes for a ranking opportunity. AI SEO embeds you within the model’s default competitive landscape.
8. Source Blending Strategy
In AI search, a “SERP” is a blend of diverse sources, not just a page.
Semrush and others note that AI engines pull from a wide range of sources, favoring community and documentation in many sectors.
AI SEO-only move:
Develop your presence into an ecosystem, beyond a single website.
Identify which non-Google platforms in your niche influence LLMs and establish credibility there.
Use consistent terminology to form a coherent online identity.
Your goal is corpus optimization, not just ranking in an index.
9. LLM-Friendly Specification Publishing
Models excel at snapping structures into place.
Content rich with detailed structures like definitions, lists, and stepwise instructions performs best in AI responses.
AI SEO-only move:
Share your key frameworks as open specifications.
Convert ambiguous messaging into clear decision-making instruments.
Document methodologies in public, thorough formats.
Offer the model a blueprint beyond just marketing speak.
10. Training-Surface Expansion
AI SEO is emerging as an industry on its own, backed by significant future investments.
However, this investment is not focused on just one index.
AI SEO-only move:
Explore potential training surfaces within your specialty like open datasets and public reports.
Place your best insights there openly, ready for retrieval or training.
Treat every public snippet as training material, not only lead generation.
You are determining where and how models will encounter your reality.
11. Anti-Hallucination Engineering
Hallucination in AI isn’t hypothetical.
Benchmarking and academic studies consistently show that AI can produce false details, particularly in low-coverage or vague topics.
AI SEO-only move:
Distribute concise fact sheets about your entity across neutral sources.
Remove contradictory public claims wherever possible.
Monitor and adjust how AI systems portray your brand.
While eliminating hallucinations is impossible, you can ensure the model opts for a well-documented version of you.
12. Mention vs. Citation Optimization
In AI searches, there are three distinct states:
Your brand is not mentioned.
Your brand is mentioned, without citation.
Your brand is both mentioned and cited.
Research indicates that citation patterns relate closely to specific quality signals on the page and sites.
AI SEO-only move:
Design pages that meet both narrative and citation criteria.
Grow earned media allowing third-party sites to be cited.
Map your current state across engines and craft campaigns to elevate your position.
Just as traditional SEO distinguishes between impressions and clicks, AI SEO separates mentions from citations, and this is crucial for visibility.
The Uncomfortable Balance
We must face some key truths:
AI summaries are raising zero-click behavior, compressing publisher traffic, with click-through rate declines between 15 to 80% depending on the query.
Platforms claim higher quality clicks and satisfaction while expanding these features into search.
Despite advances, LLMs still hallucinate, reducing errors involves better grounding and evaluation.
As individual brands, we cannot change these broad issues. But we can adapt to the current landscape:
Treat AI answers not as a novelty added to SEO but as a unique channel.
View AI SEO as a standalone channel with specific levers, measurements, and content styles.
Create content for retrieval, trustworthiness, and reuse by generative systems.
Traditional SEO isn’t obsolete, but it is only part of the journey now.
As someone deeply invested in the fascinating world of agentic commerce, I’ve become curious about what really boosts product visibility in the AI shopping realm. It’s a topic worth diving into as AI rapidly transforms the way consumers make purchasing decisions.
Have you ever wondered how platforms like ChatGPT, Perplexity, and Rufus determine which products grace the digital shelves? Uncovering this process offers valuable insights into AI decision-making and gives us a competitive edge in this new era of shopping.
Let me share with you how these AI platforms evaluate and choose products, allowing us to strategically position our offerings and maximize their AI shelf presence. Understanding these dynamics empowers us to navigate and excel in AI-driven marketplaces effectively.
Discover how I navigated the world of SEO amidst the rising influence of AI in search, with data-backed insights that show where our efforts truly pay off.
As the integration of AI into search accelerates, I’ve noticed a wave of anxiety sweeping through our community. It’s becoming harder than ever to separate robust strategies from distracting noise.
I personally found a gem of insight at the News & Editorial SEO Summit (NESS) held online last October. This article aims to provide clarity amidst the hype.
I’ve distilled the wisdom from technical SEO experts from The New York Times, Polemic Digital, and NewzDash into five actionable truths. These insights form a robust framework for our 2026 SEO strategy.
1. AI Overviews Aren’t Devouring Breaking News (Yet)
Despite the buzz that AI Overviews dominate the SERPs, data shows a much more nuanced reality for timely content.
NewzDash’s analysis reveals that only 1.9% of trending news keywords trigger AI Overviews. During high-traffic moments, traditional search features still hold their ground.
The surprising truth is that AI Overviews tend to appear several hours after news breaks, once real-time coverage slows.
Low-competition queries.
General searches without qualifiers (e.g., ‘Chicago’).
Topics lacking immediate coverage.
It’s clear: Top Stories continue to dominate when it matters, despite the AI narratives.
2. Your ‘Good Enough’ Core Web Vitals Are Probably Fine
The obsession with perfect Core Web Vitals scores isn’t necessary for solid rankings. While they indicate usability, their direct impact is less significant than many believe.
The real harm of poor Web Vitals lies in how they affect user behavior, sending negative signals to Google.
Industry experts advise reaching ‘Mostly green’ scores, investing further often yields minimal SEO returns.
3. Stop Obsessing Over Clean Code – Focus on Meaning
I’ve learned that Google’s crawlers prioritize meaningful content over spotless code. The time spent on code cleanup is better invested in enhancing semantic markup.
Using semantic HTML tags (like <article> and <section>) makes your content more interpretable to Google.
4. Content Chunking Is Your New AI Superpower
In the age of AI, structuring content for comprehension is crucial. Content chunking becomes vital for AI system visibility.
By implementing clear, logical structures with semantic HTML, we increase our odds of making our content AI-accessible.
5. Don’t Fall for the Latest AI SEO Gimmicks
Amidst the AI scramble, many gimmicks arise that don’t prove effective. It’s crucial to focus on fundamental strategies rather than chasing trends.
AI visibility remains grounded in traditional SEO practices. Google Search still underpins AI systems like Gemini AI Overview.
The Real Path to AI Visibility Still Runs Through Core SEO
As AI reshapes search, staying anchored in the basics of effective SEO is paramount. A strong foundation in technical and user-centric practices remains essential.
Ultimately, our goal is to create content that both humans and AI find undeniable, ensuring the enduring relevance of our strategies.
Why the web as we know it may fade and what AI, personal agents, and data interfaces mean for publishers, SEO, and commerce.
Every day, I’m witnessing people turn to AI for answers, product comparisons, and making quick decisions.
This shift reveals a core issue: the structure of the web wasn’t originally meant for machines.
As AI agents evolve, the way information is delivered – and the need for traditional webpages – could see dramatic changes.
The idea that the web as we know it could end, which I mentioned during a live OXD podcast in Salzburg, drew reactions ranging from thoughtful to angry.
Someone even insisted, “The web will always be there.”
Yet, those of us paying attention understand that “always” and “never” are shaky concepts in technology.
Technological history illustrates that nothing is forever.
Disruptions are noticeable only in hindsight.
Recall August 6, 1991 – could anyone foresee how Tim Berners-Lee’s World Wide Web would transform the world?
This cycle of dismissing new technology as too expensive or complex is as old as technology itself.
People pointed to existing solutions and assumed they’d last.
We also tend to judge new technologies prematurely, comparing immature models to systems we’ve heavily relied upon.
What we often fail to do is envisage the evolved state of a new technology.
This tendency clouds our future outlook.
When I’m in the market for a smartwatch, where do I usually turn for information?
Most often, I start with Google, landing on manufacturer or retailer pages.
Trying to compare the Samsung Galaxy Watch8, Classic, and Ultra to determine if the price difference makes sense is challenging.
Can I get this clarity from Samsung’s site? Probably not.
Each product page praises its uniqueness.
This forces me to jot down notes just to make basic comparisons.
I ponder over the difference between various bands and processors.
To grasp certain features, translations are sometimes necessary.
Even the “compare” function often leaves more questions than answers.
And while expectations would assume the premium model to have a specific feature, marketing priorities often arc differently.
The websites prompt more head-scratchers: Do these technical terms even matter to me?
My search broadens, throwing me onto SEO-crafted pages.
These sites often try leading me towards affiliate links.
Time is the thief here; Google requires nuanced search phrases and countless clicks.
But when I ask ChatGPT, the answer is swift and spot-on.
In less than four seconds, I get a clear comparison, making sense of all distinctions.
Follow-up questions are met with clarity.
If there are specifics to check, I am advised accordingly.
Such instances highlight the inefficiencies of web research.
Manufacturers tend to showcase products as they envision them.
But we often want straightforward comparisons.
We thrive on differences; we’re delta thinkers.
Sellers often prefer presenting products singularly.
If something isn’t present, obfuscation is the strategy.
It’s understandable, but not helpful.
Stop for a moment and try your AI for search queries.
If it’s been a while, you’re likely to be amazed.
In mere seconds, you get detailed answers.
Unsure about source reliability? Tailor your queries:
– “Only search designated expert sites.”
– “Only use well-known institutions.”
– “Give me all sources.”
The updated Google’s Gemini can produce extensive reports after an in-depth research request.
Imagine rich responses, often more comprehensive than solo human efforts.
That’s the growing strength of AI.
Using HTML makes content flexible for human consumption.
This system assists us in seeing and reading what’s online.
However, as AI usage expands, the limitations become apparent.
For example, the figures on a webpage may be clear to us, but the HTML lacks inherent semantic meaning for machines.
Structured data came as a solution but remains underused.
This impedes machine comprehension.
Apart from internal systems or large enterprises, structured data implementation is sparse.
Therefore, the primary content is still somewhat elusive to machines.
Google has worked hard to bridge this understanding gap.
Yet, AI continues to evolve, seeking innovative ways to parse and utilize data.
While AI presently gleans information through pattern matching, its potential remains vast.
Chatbots like ChatGPT offer solutions today.
The real challenge is context comprehension, which remains elusive for AI.
While both amazing and rapid, AI’s journey is just beginning.
The advances have sparked immense growth and excitement.
This era has only begun, opening doors to boundless possibilities.
Imagine a world transformed by personalized AI assistants.
The possibilities intrigue me.
These personal agents will tackle our daily routines, searching for optimal solutions.
AI might soon handle appointments, emails, and much more, offering efficiency and convenience.
Such shifts might alter how we interact digitally.
Content delivery and decision-making will evolve over time.
Our current HTML limitations challenge technological adaptability.
A new paradigm could include AIDIs assisting us with data retrieval.
Incorporating AIDIs means transitioning from HTML to structured forms.
Imagine AIDI extensions making data interpretation effortless.
Personal agents would operate even more efficiently.
The transition hinges on AI development and adoption.
Comparatively, the idea seems vast – but technological evolution often brings surprises.
Before long, our interactions may become distinctly AI-driven.
Offering a personalized touch, these agents may surpass our expectations.
I often reflect on the evolving landscape of search and how tools like Google Search and AI platforms such as ChatGPT are reshaping how we discover content. With these shifts, I’ve learned how crucial it is to track, optimize, and convert customers effectively across both platforms.
Recent developments like AI Overviews, ChatGPT, and zero-click results have led many to speculate about the end of SEO. However, I believe SEO is far from dead – in fact, it might be more vibrant than ever.
Search engines are still responsible for about 88% of all search traffic, while AI usage is nearly doubling. This dual rise tells me that consumers aren’t just choosing between Google and ChatGPT – they’re using both together.
The narrative that we must choose between SEO or AI search can be misleading. I see them as parallel paths of discovery that need to be mastered together.
People like certainty and often look to focus resources on either a tried-and-true channel or explore a new one. Yet, I’ve realized overindexing in AI while ignoring classic SEO forfeits current market share, and hesitating gives competitors a head start.
The assumption that AI growth reduces Google usage is flawed. While Google’s share fell to 89.62%, ChatGPT’s user base is soaring. Yet, from where I stand, consumers aren’t leaving Google – they are just using more platforms.
From my perspective, ChatGPT adoption has led to increased usage of Google, with sessions rising from 10.5 to 12.6 sessions per week. AI complements traditional search, enhancing the scope of our discovery process.
This expansion in search activity presents a ripe opportunity for ecommerce. Remarkably, 43% of ecommerce traffic comes from Google’s organic search, and organic traffic supports 23.6% of all ecommerce sales. Meanwhile, shopping inquiries in ChatGPT grew from 7.8% to 9.8% in the first half of the year.
The total addressable market for search visibility has multiplied, with searches now distributed across various channels. I ask myself how brands can capture this holistic search opportunity.
Tracking is essential. Implementing comprehensive tracking allows me to see the full picture of our search performance. This often requires managing traditional search statistics separately from AI results, yet the integration of tools like Semrush Enterprise AIO has been invaluable for tracking visibility across different platforms.
On the content side, key SEO principles support AI search performance, but the structure might need tweaks for optimal topical coverage. I always ask if my content answers users’ actual questions effectively. Covering vital questions upfront boosts relevance and the potential for AI citation.
Giving content full context is another principle I adhere to. AI models view topics as connected ideas. Writing about sustainable products means also discussing eco-friendly materials and related subtopics, but without resorting to keyword stuffing.
Ensuring my content is accessible to both AI and humans means prioritizing readability, clarity, and logical structure. It means everything from heading hierarchy to scannable formatting must be on point.
Platforms like Semrush Enterprise AIO help by offering dual-channel optimization capabilities that I find reduce guesswork and provide guidance for maximizing search performance.
Profit is the ultimate focus, and I’ve found that AI search visitors are 4.4 times as valuable in terms of conversion. Coupling this with search engines’ role in brand discovery shows the importance of optimizing across both avenues.
To me, the outdated choice between SEO and AI is a misunderstanding of modern search discovery. Customers aren’t choosing – they use both Google and ChatGPT, often simultaneously.
By embracing this dual-channel approach, brands are poised to dominate the search landscape, ensuring they are present wherever customers begin their search journey.
