Are Citations In AI Search Affected By Google Natural Visibility Modifications?

I recently wrote about an unconfirmed Google algorithm replace that rolled out in mid-January 2026, which negatively impacted the natural search visibility of dozens of main manufacturers. For many of the impacted websites I analyzed, the influence was disproportionately focused to the firm’s weblog, or one other folder containing informational articles and sources.

That very same natural trajectory has continued into mid-February for all of the subfolders I analyzed, utilizing the Sistrix U.S. Visibility Index:

Zooming out, this is what the drops seem like if you take a look at the visibility developments throughout the entire domains, not simply the blogs:

Right here is one other instance of the visibility influence on the firm weblog for the largest firm in the record (when it comes to each ARR and natural visibility):

And this is what the influence appears to be like like if you take a look at the firm’s full area’s visibility in natural search:

Pointless to say, these latest natural visibility drops had been excessive, relative to the websites’ total search engine optimisation trajectories over the previous few years. Drilling down into 11 of the websites that noticed excessive declines over the final month, I wished to see if this new knowledge may assist reply one other query:

Do drops in Google natural search visibility coincide with related drops in AI search citations?

My working speculation is that these drops are not simply remoted to conventional search. As an alternative, I think we’ll discover that, for many LLMs, AI search quotation developments mirror what occurs in Google’s natural search outcomes, for 2 causes:

1. The Direct Pipeline: Google’s AI Ecosystem

For Google’s personal AI merchandise – AI Mode and Gemini – the correlation needs to be strongest. Presumably, Google is utilizing its personal index and top-ranking search outcomes to formulate AI search responses; due to this fact, dropping in natural rankings ought to logically trigger these pages to be cited and referenced much less regularly in generative solutions.

2. The Downstream Results: Third-Get together LLMs (ChatGPT & Perplexity)

The hyperlink between Google natural rankings and third-party LLMs like ChatGPT and Perplexity is extra nuanced, as we don’t know precisely which serps these LLMs are surfacing for internet search.

Whereas there is a growing body of evidence (and industry reporting) suggesting that ChatGPT probably scrapes Google throughout dwell internet searches, we nonetheless technically lack official affirmation from the supply. Perplexity, on the different hand, is currently believed to utilize the Brave Search API as a core a part of its retrieval course of, alongside its personal specialised “PerplexityBot” crawler.

To check this out, I wished to drill down into the subfolders that noticed substantial visibility drops on Google in latest weeks, to see whether or not the development line for AI search citations adopted swimsuit.

To begin, I honed in on an inventory of 11 websites whose subfolders noticed substantial natural visitors drops between January 20, 2026 and February 16, 2026.

I used the Ahrefs MCP server with Claude Cowork to pull in estimated world month-to-month natural visitors numbers for every path (subfolder) in the record. As a result of most of the visitors declines began round January 21, 2026, I pulled the projected month-to-month natural numbers for January 20, 2026 and the most up-to-date date, February 16, 2026.

I additionally redacted the website names, leaving the title of the subfolder and a short, anonymized abstract about the firm kind and the subfolder’s objective:

These subfolders skilled anyplace from a -5.7% to -53.1% drop in estimated month-to-month natural search visitors since January twentieth, 2026.

Utilizing Ahrefs Model Radar, you possibly can drill down to see the variety of AI search citations {that a} given subfolder has obtained throughout varied LLMs over time. For instance, right here is the ChatGPT quotation development line for the first subfolder listed in the above desk (U.S. knowledge):

This is the corresponding chart displaying the natural visitors development for this similar subfolder, which started dropping round January 21, 2026:

I used the Ahrefs MCP server with Claude Cowork to pull world visitors and quotation knowledge, and to analyze this similar sample for 11 of the subfolders that noticed massive drops.

Observe on methodology: Whereas 11 subfolders are a small pattern measurement, I used to be particularly searching for a “clear” knowledge set – subfolders experiencing the same algorithmic demotion on Google throughout the unconfirmed January 2026 replace. By narrowing the scope, I may higher isolate whether or not a lack of conventional search visibility interprets straight right into a quotation drop in AI search.

Beneath are the high-level summaries of how natural visitors and quotation counts modified throughout Google and varied LLMs, together with AI Mode, ChatGPT, Perplexity, and Gemini:

Findings:

• The information exhibits a broad decline in each search engine optimisation visitors & AI search citations: Each subfolder in the examine (11 of 11) skilled a drop in each Google natural visitors and whole AI search citations, with a big common quotation decline of -22.5%.
• Google’s AI Mode (-23.8%) and ChatGPT (-27.8%) confirmed the most extreme declines, carefully mirroring the -26.7% common drop in natural visitors.
• Whereas Gemini additionally noticed broad declines (10 of 11 websites), Perplexity proved to be the most resilient, with solely 4 of the 11 websites seeing a drop and a a lot milder common change of -2.9%.
  • This knowledge helps the principle that Perplexity is primarily utilizing non-Google search surfaces to generate its responses.

Taking a look at the modifications in estimated natural visitors for every subfolder in contrast to whole AI search citations between January 20 and February 16, 2026, the correlation is clear: Important losses in natural search visibility are nearly universally mirrored by a corresponding decline in AI search citations.

Drilling down into particular LLMs, together with Google’s AI Mode, ChatGPT, Perplexity, and Gemini, exhibits how the decline was practically common for many platforms, whereas Perplexity regularly displayed a big divergence, displaying optimistic quotation development for the majority of the subfolders regardless of their natural visitors losses.

ChatGPT (inexperienced) persistently exhibits the deepest declines throughout nearly each subfolder – usually exceeding AI Mode and Gemini. This is intriguing as a result of ChatGPT isn’t a Google product, but it seems extra delicate to these natural rating shifts than Google’s personal Gemini.

This seems to be one other clue that ChatGPT is reliant on Google’s search index throughout retrieval.

AI Mode and Gemini have a tendency to transfer in the similar path however not the similar magnitude. Regardless of each being Google merchandise, AI Mode declines are usually steeper than Gemini’s. This may counsel they weight or supply from Google’s natural index in a different way – maybe AI Mode is extra tightly coupled to dwell SERP rankings whereas Gemini attracts from a broader or cached information base.

The few websites the place Perplexity did decline (e.g., Website J, Website Okay) are additionally the ones displaying comparatively smaller natural drops. So even in the circumstances the place Perplexity tracked downward, it doesn’t seem to be correlated with the severity of the Google natural loss – additional proof that Perplexity is probably pulling from a special retrieval pipeline.

The beneath desk exhibits all the natural search vs. AI search quotation knowledge in a single place:

The desk reveals a transparent sample: each subfolder that misplaced natural visibility on Google additionally noticed a decline in whole AI search citations, with a median drop of -22.5% throughout all LLMs.

ChatGPT was the most severely impacted platform, with quotation declines reaching as excessive as -42.3% (Website E) and exceeding -34% for 5 of the eleven subfolders – usually surpassing even the natural visitors loss itself.

Google’s AI Mode adopted the same trajectory, whereas Gemini confirmed extra average declines throughout the board.

Essentially the most notable outlier is Perplexity, which really confirmed quotation development for 7 of the 11 subfolders, reinforcing the principle that it retrieves from a non-Google search index.

Maybe the most fascinating discovering is that ChatGPT – a non-Google product – seems extra tightly coupled to Google’s natural rankings than Google’s personal Gemini, suggesting that ChatGPT’s internet retrieval pipeline is closely dependent on Google’s search outcomes.

One suggestion I’ve been making since AI search entered the search engine optimisation dialog is that you shouldn’t spend money on AEO/GEO techniques that might be detrimental to search engine optimisation efficiency. For instance, utilizing hidden immediate injections, cloaking, or self-promotional listicles (techniques that some have advocated for to increase AI search visibility) is likely to be quickly useful for AI search, however may trigger huge complications with Google and Bing’s natural search rating algorithms down the line.

Now, now we have much more proof that AI search is essentially linked to search engine optimisation efficiency: In the event you drop in natural search, you possibly can probably anticipate a corresponding drop in citations not solely from Google’s personal AI search merchandise, however from different LLMs like ChatGPT, which seem to even be closely reliant on Google’s search outcomes.

The one notable exception is Perplexity, which confirmed quotation development for the majority of subfolders hit by the latest algorithm replace. That stated, it’s essential to weigh this in opposition to the scale of visitors and LLM utilization at stake. In accordance to a recent article by Similarweb, ChatGPT obtained 5.8 billion internet visits in August 2025, in contrast to 148.2 million for Perplexity.

So as to add to this, if you think about Google’s natural search visitors – which nonetheless dwarfs all the AI search platforms mixed – the overwhelming majority of your search-driven visibility throughout each serps and AI chatbots is nonetheless flowing by way of a pipeline the place Google’s rankings dictate the end result.

For the previous 12 months, the search engine optimisation trade has been asking how carefully traditional SEO and AEO/GEO are actually tied collectively. I believe this knowledge helps reply that query: Not solely is a robust search engine optimisation basis important for AI search visibility, however techniques that harm your natural rankings can have a cascading unfavorable influence on your AI search citations as nicely. In different phrases, the quickest manner to lose visibility in AI search is likely to be to lose it in Google first.

Extra Assets:

This publish was initially revealed on Lily Ray NYC Substack.

Featured Picture: PeopleImages/Shutterstock