In Boston’s Seaport District, the halls of the Westin were buzzing with a mixture of excitement and apprehension as industry leaders gathered for SMX Advanced. As the first in-person SMX conference in years, this was a major event for attendees and crew alike.
What emerged from the conference wasn’t just another set of tactical updates or algorithm tweaks, but a collective recognition that as an industry, we’re witnessing the most significant transformation in SEO in all of our careers.
From Deterministic to Probabilistic: the New Search Reality
Dawn Anderson said it best, declaring what many attendees were already feeling in their bones:
“This actually represents the biggest shift in information retrieval in 70 years.”
Dawn Anderson
The traditional paradigm of index, retrieve, and rank is giving way to what she called, “a converged, single model”, driven by generative AI.
This shift from deterministic to probabilistic search was echoed throughout the conference. Mike King elaborated on this fundamental change:
“The main thing that you need to know about the change here is that we’ve gone from this deterministic environment to a probabilistic environment… Now, we’re in this probabilistic environment where you can put the same content in multiple times, and it may come out different. So, you have a lot less control than you ever did.”
Mike King
The implications are staggering. Where SEO once operated on semi-predictable cause-and-effect relationships, we now face an environment where the same input can yield different outputs each time.
What This Means for SEO
SEOs need to start relying a little less on what keyword research tells us and what competitors are writing about. Instead, we need to be getting into the minds of our target audiences, understanding every possible barrier to entry, and creating content aimed at addressing those barriers.
In the age of AI customizing unique answers to complex questions, having content that addresses every facet of search will help increase brand visibility and awareness.
The Great Decoupling: Traffic Patterns in Flux
Mike King discussed the concept of the “great decoupling,” a trend where impressions increase while clicks decline. He shared data to back this up, stating that, “Wikipedia’s traffic continued to grow after ChatGPT’s rollout, but then saw a significant downward trend when AI algorithms started to appear.”
This isn’t isolated to Wikipedia. King revealed that even HubSpot, despite being an inbound marketing pioneer, saw dramatic traffic drops coinciding with the AI Overviews launch. Yet paradoxically, their revenue continues to climb, suggesting that, as King stated, “AI overviews are sending more qualified traffic when people click through.”
Mordy Oberstein added another dimension to this traffic discussion, revealing that branded traffic has declined by 11%, with his data team suggesting simply, “People are just sick of brands.”
That begs the question: are people really sick of brands? Or are people just searching differently?
What This Means for SEO
This fracture of impressions and clicks has been seen and heavily reported on by the SEO industry. However, we should probably pump the brakes before restructuring our entire KPI set.
The reality is this: we don’t know how users are actually leveraging AI in search. Talk to your non-technical friends to see how they’re using it. You’ll find that most of them know a lot less than the SEO community assumes they do.
It’s entirely possible that society rejects the overuse of AI due to mistrust. However, that may not stop Google from attempting to force it on people.
So, can we really say that people are sick of brands? Probably not. Good branding is what helps companies last. It’s what brings back repeat customers. Let’s also not forget that browser autofill is a thing, so return customers are more likely to use that than to search in Google.
The Rise of Relevance Engineering
Another interesting concept introduced was Mike King’s vision of “relevance engineering”: a convergence of information retrieval, artificial intelligence, content strategy, and digital PR. He challenged the audience with the following:
“SEO is not prepared for any of this. Everyone in our space… are saying that it’s just SEO. This is a paradigm shift in how this stuff works… This is an opportunity for us to reset what our industry means, what the value is, and how we do things.”
Mike King
King’s point was clear: clinging to traditional SEO practices in this new environment is not just outdated, it’s actively harmful to success. “We are no longer just optimizing,” he says. “We are engineering.”
What This Means for SEO
This ideology represents a micro fracture in the SEO community. Many take Mike’s side of saying it’s not just SEO anymore, while others claim that SEO is a practice and profession and we must simply adapt to the everchanging search environment.
Mike is correct that we will need to adapt to how Google and other AI chat providers are changing the landscape. However, it’s up to you to decide if you want to adopt another form of initialism for your job title.
Where might we end up? Likely using our traditional SEO skills and tactics, but evolving to use them in different ways. The SEOs who will succeed in this new world of search are the ones with refined forensic skills. Meaning, if you can understand how to reverse engineer the way results appear in different AI-generated content, you might be able to replicate success.
The Agentic Future is Already Here
Dave Davies painted a picture of the near future that’s already beginning to materialize. One where AI agents act as intermediaries between users and information. He emphasized the fundamental shift in user behavior:
“Agents will increasingly act on behalf of users, necessitating a re-evaluation of optimization strategies… marketers will need to optimize for these agents, rather than solely for human users, as agents will be navigating sites and making decisions for them.”
Dave Davies
Davies predicts that within one to two years, marketers will be actively optimizing “agent cards” – a new frontier in digital marketing that will define how agents communicate with each other and make decisions on behalf of users.
What This Means for SEO
This is a very interesting take, and one that may separate the technical SEOs from the content marketers.
This author’s take is that AI is trying to replicate what humans want. So if we’re optimizing for humans, then we’re inherently optimizing for AI. However, understanding information discovery and retrieval may help us better understand how to make our content easy for AI agents to use.
Brand as the Silent Killer of Performance
In a particularly interesting session, Mordy Oberstein introduced the concept of “brand decay” as, “the silent killer of performance.” He argued that misalignment between brand messaging and actual user experience is creating a cascading effect on all performance metrics.
His warning signs of brand decay included:
- Mismatch between brand messaging and natural audience
- Inauthentic social media engagement
- Lingering reputation issues
- User queries like, “Should I trust [Brand]?”
The message was clear: in an AI-driven search landscape, brand coherence and authenticity aren’t just nice-to-haves. They’re actually survival mechanisms.
What This Means for SEO
It’s true, brand is important. But inconsistency in messaging, tone, and theme can create a disconnect between the customer/company relationship. This is beyond SEO and more of a fundamental marketing principle.
Having misalignment in content messaging can decrease the user experience and may lead to LLMs having a misconception about your brand.
The Human Element: Distribution and Authenticity
Wil Reynolds delivered perhaps the most passionate plea of the conference, challenging attendees to move beyond algorithmic thinking to focus on human behavior and authentic content creation. His criticism of the industry was biting:
“You create your own trust, you create your own brand by copying other people. How about we start changing that?”
Wil Reynolds
Reynolds introduced a powerful metric for content quality: “If it gets searched but [gets] no social [interactions], it means you know why you love it… Traffic from social media is a ‘human vote.’” He revealed that his own content generates the same amount of conversions from social as from search, despite social delivering only one-tenth the traffic.
What This Means for SEO
Many folks have talked about traffic source diversification. But Wil gave us a sobering reality check that it’s not about simply acquiring traffic from other sources. It’s about creating content and a brand that people want to engage with.
If you truly understand your target audience, their challenges, and how they engage with your niche on social, you can tailor your content to be share-worthy.
The MoE Revolution: Understanding How AI Thinks
Crystal Carter introduced the audience to Mixture of Experts (MoE) models, explaining how these new AI architectures fundamentally change how queries are processed. Unlike older dense models, MoE systems route queries through specialized “experts” selected from a diverse bank of skills.
Her practical advice was this: “Treat each query like a keyword,” and understand that when search is integrated, MoE models “break down queries, look at search results, and activate websites as ‘experts.’” This represents a fundamental shift in how content needs to be structured and optimized.
What This Means for SEO
Topical authority will be a major factor in deciding whether or not Google and other LLMs will show your brand for certain queries. This is not a new concept, but more of a double down on creating content that reinforces your core competencies.
Measurement in Crisis: the Tools Aren’t Ready
A recurring theme throughout the conference was the inadequacy of current measurement tools. Mike King didn’t mince words: “SEO software in general is not ready for this moment.”
Jess Scholz reinforced this, noting that with personalized AI-driven results, “conventional tracking methods lose their reliability.” The challenge is compounded by Google’s limited transparency. As King noted, while AI overviews and AI mode data are supposedly in Google Search Console, “you can’t filter for it.”
What This Means for SEO
This does mean that we’re going to be having some tough conversations with stakeholders. But that doesn’t mean you need to throw out your dashboards and cancel your subscriptions. SEO tools are actively deploying tactics to attempt to track visibility in AI Overview. While it’s not perfect, it’s a start and it’s better than nothing.
This author’s challenge to the community is that instead of criticizing the validity of tools, attempt to try something new. Our industry only moves forward thanks to the brave people that try.
The brilliant team at LOCOMOTIVE has been actively working towards groundbreaking methodologies for tracking performance in LLMs. We urge you to take the time to experiment and try to create something new for the industry. We’re a community and we only grow together.
Practical Strategies for the New Era
Despite the challenges, speakers offered concrete strategies for navigating this new landscape:
1. Embrace Semantic Clarity
Dawn Anderson emphasized the need for consistency across all content surfaces, avoiding ambiguity, and implementing, “topic chunking in long form content.”
2. Focus on Distribution
Wil Reynolds’ data showed that content with strong social distribution signals generates significantly higher conversion rates, even with less traffic.
3. Build for Multiple Contexts
Crystal Carter advised creating content that addresses nuanced, specific questions about products and services, going beyond general information to become the definitive expert source.
4. Production Speed Matters
Jess Scholz stressed that, “every single minute that your content is not in, is a conversion lost,” advocating for API usage and multiple indexing strategies.
5. Think Beyond Your Website
Multiple speakers emphasized that optimization now extends far beyond your own domain, encompassing YouTube, social platforms, and emerging AI interfaces.
Looking Forward: Evolution, Not Extinction
As the conference drew to a close, a consensus emerged: while the challenges are real and the changes profound, this represents an evolution, not an extinction event for digital marketers. Dawn Anderson captured this sentiment perfectly:
“It’s not the end of classic information retrieval. For us, classic search is just a new and exciting time.”
Dawn Anderson
The message from SMX Advanced was clear: those who cling to outdated practices will be left behind, but those who embrace this new paradigm and who understand that they’re no longer just optimizing but engineering relevance across a complex, probabilistic landscape, will find unprecedented opportunities.
As Mike King concluded in his session: “The future is what we decided it is. We don’t have to be beholden to what SEO has been for the last 25 years. We can do something different.”
The question for every digital marketer leaving Boston wasn’t whether to adapt, but how quickly they could transform their practices for this new reality. The seismic shift in search isn’t coming. It’s already here.
LOCOMOTIVE is committed to staying at the forefront of these industry changes, helping our clients navigate the evolving search landscape with strategies that balance technical excellence with authentic human connection.
