AI Attacks! SEO and LLMs Collide
The world of SEO has never been one of concrete ideas and exact sciences. While SEO experts have developed numerous reliable ways to deliver results for clients, the targets have always been shifting as Google and other search engines changed their algorithms.
Then, in June 2025, everything changed when Google dropped its biggest update yet. The idea of this update, per Google, was “designed to better surface relevant, satisfying content for searches from all types of sites.” Seems simple enough, right? Well, this was no simple update; it was Google’s most significant push to incorporate its newly developed Large Language Model (LLM) systems into its search engine, in an effort to stem the expected loss of search revenue to popular LLMs like ChatGPT.
In a matter of days, many clients of ours had their SEO rankings completely tank, some seemingly losing all progress their SEO campaigns had previously made. It was a rough few weeks, to be sure, and quite an unnerving time for our agency. But, rather than panic, our team dug in. We began monitoring what other SEO specialists in our industry were doing, the data they were seeing, and their thoughts and predictions about how SEO would operate going forward. We even went to Austin, TX, to attend the SEO conference PubCon, and got to talk with a lot of people in our industry who were experiencing the exact same issues we were. Being in a room full of SEO experts, it was clear to see that this wasn’t one of Google’s normal algorithm updates.
Where’d All the Clicks Go?
Ultimately, with the new update, several ways it impacted search severely hampered the ability to properly implement and track SEO, the most significant being the rise of zero-click behavior. Because Google’s AI Overviews often gave users the answers they were seeking, without ever having to scroll down the Search Engine Results Page (SERP), the websites that worked so hard to rank top of page one were no longer being clicked. Even with a page ranking #1 on a Google SERP, it was producing vastly fewer clicks.
What we soon realized was that the problem wasn’t just fewer clicks — it was that Google was now far better at understanding intent and meaning across content. Pages that once felt distinct to humans were increasingly being interpreted as saying the same thing, just with different wording.
It’s Just Semantics
This shift was driven by a major evolution in how Google understands and evaluates content. Rather than relying primarily on keyword matching, Google has become far more capable of interpreting meaning, intent, and topical overlap across pages. In practice, this means the search engine can now recognize when multiple pages are attempting to answer the same question — even if they use different wording to do so.
One example of this evolution is Google’s MUVERA system (Multi-Vector Retrieval via Fixed Dimensional Encodings), which reflects how far search has moved beyond simple keyword matching. Systems like this allow Google to evaluate meaning, intent, and contextual similarity at scale, making it far easier to recognize when multiple pages are attempting to solve the same problem — even if they use different language.
Why This Broke Our SEO Model
Once Google began answering questions directly and evaluating content based on meaning rather than keywords, the traditional SEO model began to unravel. Strategies built around maximizing clicks and keyword coverage were suddenly misaligned with how search now worked — even if they had been effective for years prior.
For example, much of our content writing strategy was built around driving clicks. By ranking highly for as many keyword variations as possible, we could more reliably drive targeted traffic to our clients’ websites through search. This was done by building multiple, uniquely written but functionally identical pages for targeting multiple keywords.
But now, with Google disincentivizing clicks and better understanding semantics, our old SEO pages were no longer uplifting our client’s visibility in the most meaningful way. On top of that, Google no longer saw the value in so many different pages with similar content, and started to de-rank content it deemed redundant. In one fell swoop, our old keyword-page strategy lost both visibility and relevance, leaving many of our clients in the lurch.
The Time for Action
However, those of us involved in SEO are no strangers to change and times of ranking volatility. This was a larger challenge than many of us had seen in SEO before, but we were ready and willing to take it on!
E-E-A-T Up!
Thankfully, it didn’t take us long to identify how we would change our SEO strategy; to overcome the new, we had to embrace the old. The idea was to utilize much of what Google had already laid out for us, in terms of its own guidelines for highly ranking content. The bridge between the new and old SEO thinking was E-E-A-T (Experience, Expertise, Authority, and Trust). Rather than build pages that incorporated elements of E-E-A-T across numerous pages, we had to hit every E-E-A-T signal in significantly fewer pages. We had to ditch the thin-but-optimized pages of the past, and begin building fewer, stronger pages that acted as the central hub for our target keywords and their variations.
This realization forced us to rethink not just what we were writing, but how our content was organized and connected across an entire site.
An SEO Foundation Built on Pillars
Soon, we developed a new SEO strategy for our clients based on primary and secondary pillars, along with additional location pages for local SEO value. In our development of this strategy, it made sense for several reasons:
- Focus Shift – Google’s move from keywords to topics required a corresponding shift in how content is structured. Instead of pages designed to signal the presence of a specific keyword, our pillar pages are built to demonstrate a complete understanding of an overarching topic or subtopic. This alignment with how Google now evaluates relevance helps ensure content is interpreted accurately across SERPs and LLMs.
- Redundancy Reduction – As Google’s semantic understanding improved, content that was technically unique but contextually similar began to lose value. By consolidating information into fewer, more comprehensive pages, we reduce overlap and ensure each page exists to add meaningful value. This makes it easier for search engines and LLMs to identify authoritative sources without wading through near-duplicate content.
- Improved Hierarchy – The pillar model creates a clear, intentional content hierarchy that mirrors how both users and search systems navigate information. Primary pillars establish topical authority, secondary pillars expand on key subtopics, and supporting pages add context where needed. This structure improves usability for visitors while helping search systems more reliably match pages to user intent.
- Increased Mentions – Strong, well-structured pillar pages that clearly demonstrate Experience, Expertise, Authority, and Trust are more likely to be referenced by AI systems when generating answers. When LLMs identify content as comprehensive and credible, they are more inclined to surface it in responses – whether through direct mentions, citations, or links – extending visibility beyond traditional clicks.
But Does It Work?
Over the last eight months, the 309 crew has had their noses to the grindstone, testing and rolling out this strategy across both new and legacy clients. Because pillar pages are more involved than traditional SEO pages, this process required extensive research, writing, and AI-assisted testing. Now, after more than half a year of focused effort, we’re beginning to see the fruits of that labor.
Even before now, we were finding that many of our new pillar pages were appearing more reliably toward the top of SERPs and consistently reflecting the proper keywords they were targeting. However, some of our biggest successes with this new strategy have come from the way our clients’ websites are now appearing more often in Google AI Overviews, Gemini, ChatGPT, and other LLMs. It’s not a perfect science, and our tracking tools haven’t yet kept pace with our methods, but manual testing of our clients’ websites has increasingly found them mentioned in LLMs, where users are asking questions relevant to their issues.
Positive First Results!
For example, one exterior renovation client came to us with the goal of overcoming the BIG DOG competitor in their community, knowing it was a lofty aim. We expressed this concern to them as well, but saw it as an opportunity to test our new strategy. Surprisingly, once our client’s website and pillar content were launched, it only took a few months for our client to go from non-existent to catching up to and even outranking their competitor for various key terms. Better yet, our client consistently showed up alongside other local contractors in Google AI Overviews.
Another client of ours, an attorney, saw similar success with their new pillar content. We redid much of the existing service pages on their website, turning them into fleshed-out informational hubs, and redirected old SEO pages to the new pillars. Several months later, the attorney began to notice that new clients had mentioned finding their firm through ChatGPT and other LLMs. All in all, these were exciting developments for our team, and encouraging that we were taking the right direction with our SEO strategies.
The Big Takeaway
So, as the dust begins to settle, what’s the lesson in all this? After navigating one of the most disruptive shifts SEO has seen in years, the takeaway wasn’t that previous strategies were wrong — it was that search itself evolved. Google’s growing use of semantic understanding and AI fundamentally changed how relevance is evaluated, narrowing the effectiveness of tactics that were once both logical and successful. In response, we’ve refined our approach toward fewer, stronger pages built to demonstrate clear authority, long-term value, and adaptability in an increasingly AI-driven search landscape.
The Future is Now
AI is here, and it’s begun to change many aspects of our lives and professions. However, rather than stomp our feet and scream “unfair,” we chose to adapt. And, it seems our gambit has paid off so far, and is leading us in the right direction.
Now, does this mean it’s all smooth sailing from here on out? Absolutely not. Google is never going to stop updating and improving its algorithms, which means SEO experts will never stop adjusting their strategies. While “big changes” like the June 2025 update won’t be all the time, there’s no telling what systems Google and LLMs decide to “upgrade” next. However, our team is more confident than ever and is ready to take the AI revolution in marketing head-on!
If you’re interested in learning more about AI or LLMs and how they’re impacting your SEO, give 309 Marketing a call today!