“If you don’t define your brand, AI will do it for you” (and trust me, you won’t like it)
Welcome back to SEO Confidential, our weekly column where we bring a little order to the chaos raging through the dazzling world of SEO and digital marketing in the age of AI. If you too have the feeling that in recent months everyone is talking about models, interfaces and miracle solutions, but less and less about strategy, you can’t afford to miss this interview.
LLMs, AI Overviews, generative content and zero-click are changing the rules, often without the data actually being read. In this conversation we tackle some of the key issues for anyone working with brands and content: why technical complexity is not the same as authority, why message consistency has become a decisive factor, how the lazy, careless use of AI is starting to weaken strategies and, above all, how to defend a brand’s reputation when AI systems reconstruct narratives by drawing on external sources, even inaccurate ones.
The guest we’ll be discussing all these topics (and much more) with is Carolyn Shelby, founder of CSHEL Search Strategies, a consulting firm specializing in advanced SEO and visibility strategies for the AI era. Active in the industry since 1994, she has worked on projects for global players such as Disney’s ESPN and for major US publishing groups, including the Los Angeles Times, Chicago Tribune and New York Daily News, and has also collaborated with Yoast SEO on the development of the world’s most widely used plugin.
If you work in business, branding or digital strategy, this interview offers concrete insights into what really matters today and what risks becoming a hidden cost tomorrow. Enjoy the read: some answers deserve to be read more than once, trust me!
“How do you defend your brand’s reputation from AI hallucinations? By leaving fewer gaps to fill!”
In recent years, SEO has often rewarded complex structures, hidden information and “technical” solutions rather than clarity. Today, with search engines and LLMs needing to truly understand what a brand offers, aren’t we discovering that much of the optimization done so far has made websites less readable rather than more authoritative?
In many cases yes, but not because clarity has stopped being important. It’s because, for a long time, shortcuts worked.
“Optimization” aimed exclusively at aggressive traffic growth pushed many teams to create pages that worked technically (for search engines) but didn’t actually communicate anything useful to human beings. Once they found success with those tactics, they went all in on filler content, overly elaborate internal linking and pages that dodged answers to inflate word counts and keywords, instead of simply stating the message or the point clearly.
What’s changing now is that both search engines and LLMs have far less tolerance for that kind of noise. They don’t reward clever concealment or artificial complexity. They reward content that gets to the point, explains things clearly and stays consistent.
So it’s not really a new direction, but rather a correction. Authority never truly came from complexity. It came from being understandable and reliable. AI just makes it harder to fake.
If the goal is no longer just rankings, but being found, cited and considered trustworthy across multiple platforms, how dangerous is it to have inconsistent messaging across websites, profiles, feeds and product pages? Is the real challenge of SEO in 2026 still technical, or is it mainly a matter of communication and strategic discipline?
Inconsistency is one of the biggest risks brands are underestimating right now.
When a website says one thing, product pages say something slightly different, social profiles tell another story and third-party sources fill in the gaps, humans might shrug. But over time, those inconsistencies pile up. The brand loses reliability and becomes harder to describe coherently.
We see a similar problem with the rise of AI-written content that uses a very recognizable cadence. The problem is not that LLMs struggle with that style… because they don’t. They created it. The problem is that human beings find it exhausting to read. It sounds synthetic and feels artificial and imprecise, especially in long-form content.
When people disengage, they stop sharing, citing and referencing that content. And it’s precisely those human behaviors that shape the third-party signals AI systems rely on. The models don’t reward good writing directly, but they absorb the world’s reaction to it.
What works is alignment and restraint. Clear language, consistent messaging and writing that respects the reader’s attention have a greater impact. When those ideas are repeated across owned channels and reinforced by credible third-party sources, they get amplified.
By 2026, SEO will still be technical in the sense that crawlability, structure and performance will matter. But those are table stakes. The harder problem is discipline: knowing who you are, saying it clearly and saying it the same way everywhere.
SEO is no longer just about being found, but about being taken seriously long enough to matter.
In many companies, AI is now used to define briefs, structure content and even shape SEO priorities. Where, in practice, does that delegation become dangerous? What signals indicate that a team has stopped using AI as support and started using it as a substitute for human judgment?
Things usually start to go wrong when AI takes the place of institutional knowledge.
Using AI to speed up research, summarize inputs or explore options is legitimate. The risk emerges when you let it define briefs, structure content or set SEO priorities without anchoring those decisions in what the organization actually knows about its customers, its products and its own history.
The warning signs are fairly obvious. One is when nobody can explain why a piece of content exists without referring to the tool. Another is when content looks polished but generic: familiar language, familiar structure, but no trace of real expertise or experience built up over time. You also see a flattening effect, where everything sounds the same regardless of audience, market or context.
In those cases, internal knowledge isn’t being applied. The model fills the gaps, but it doesn’t know what actually matters to the business, what has already failed in the past, which assumptions are risky or where nuance is needed. It can’t weigh trade-offs or remember the reasons behind earlier decisions.
AI works best when it supports people who truly know their subject. When it’s used as a substitute for that understanding, teams don’t just produce weaker content: they gradually lose the ability to reason about their own strategy.
In marketing, AI is often discussed as a tool to “scale”, far less as a matter of control, judgment and accountability. If a brand publishes incorrect, misleading or inconsistent content generated by an LLM, who really pays the price: the algorithm, or the person who decided to hand it the controls?
The brand always pays the price. The algorithm never does.
Search engines don’t lose credibility. Models don’t suffer reputational damage. Companies do, along with the people responsible for the decisions behind the content.
The “AI for scale” debate often forgets that publishing remains a choice. Letting an LLM generate or influence content doesn’t remove accountability, it concentrates it. Someone decided to trust the output, to skip a review, or to move faster than their own controls allowed.
The consequences aren’t always immediate, but they compound. Inconsistent or misleading content gets indexed, summarized, cited and reused. Customers lose trust. Journalists stop referencing the brand. Internal teams start working from fragile assumptions. None of this gets fixed with a quick edit.
AI can help you move faster, but it can’t decide what is acceptable, correct or appropriate. Those are judgment calls. When brands delegate those decisions to a system that doesn’t understand context or consequences, they aren’t scaling: they’re gambling.
Accountability doesn’t disappear when AI enters the picture. It becomes harder to avoid.
Today, as you’ve written, a lot of online content adopts a hyper-dramatic, fragmented, performative style, designed more to capture attention than to inform. Are we witnessing the birth of a new form of “keyword stuffing”, this time stylistic, that risks eroding trust and credibility over the long term?
No, it’s not really comparable to keyword stuffing. It’s something different, above all in who it hurts.
Keyword stuffing failed because it tried to manipulate algorithms, which eventually fought back. Here the problem is different: the offense isn’t aimed at machines, but at readers.
Readers immediately recognize the typical AI cadence, especially in long texts, and they read it as a sign of carelessness. They don’t just think AI was used, but that nobody took responsibility for shaping, refining or even carefully reading the text before publication.
LLMs are indifferent to that style because they generate it themselves. Human beings are not. They turn away because the writing feels generic and imprecise, and because it suggests that speed and scale were prioritized over clarity and accountability.
It’s not a new form of keyword stuffing. It’s a crisis of editorial judgment. Brands aren’t being penalized by machines, but judged by readers, who stop believing that a real person bothered to write or review what they’re reading. Once you slide down that slope, credibility erodes fast.
Carolyn, many people talk about “optimizing for AI” as if it were enough to adapt content or chase the model of the moment. But if visibility increasingly depends on technical structure, real authority and brand consistency, aren’t we underestimating the fact that AI doesn’t reward tricks, but amplifies those who have already built a solid system in the real world?
Yes, and it’s a misunderstanding that sits at the root of a lot of bad decisions.
“Optimizing for AI” gets treated as something to bolt on afterwards, or as a set of tactics to chase as the models change. In practice, that strategy fails almost immediately. Models evolve at an aggressive pace, interfaces are rewritten in real time. Any new tactic rarely lasts a quarter, sometimes not even a month.
While everything is being rewritten by AI in real time, humans keep designing websites built on what we might call “digital jazz hands“: heavy animations, layered interactions, staged reveals, glitter for its own sake. They work in design reviews, but they almost always make content harder to access, interpret and reuse, for machines and for users, who are already drifting away from the web as an interface anyway.
LLMs don’t see the show. They absorb structure, language and message. When key information is hidden behind interactions, fragmented across components or wrapped in excessive presentation, it becomes fragile. If the content can’t be extracted cleanly, the beauty of the experience doesn’t count.
What endures is structure, authority and consistency. If a website is technically accessible, if content is clearly organized, if the brand has a consistent voice and a real presence beyond its own site, AI systems have stable points to anchor to. Without those elements, no amount of stylistic care or model-chasing produces lasting effects.
AI doesn’t reward clever hacks. It reflects and amplifies what already exists in the real world. Brands that have invested in expertise, clarity and consistent repetition surface more often because the signals reinforce each other. Those that haven’t built that system tend to fade away as the ground keeps shifting.
The apocalyptic narrative about the death of SEO has spread rapidly, often at odds with data showing a moderate decline in traffic and a substantially stable structure of search. Isn’t there a risk that this alarmism will push companies and CMOs to cut organic investment precisely when brand authority is becoming even more decisive in the AI era?
Yes, and it’s a very real risk.
It’s striking how disconnected the narrative is from the data. In most industries we’re not seeing a collapse of search, but a rebalancing. Traffic is shifting, not disappearing. The real collapses involve models built on scale without value, such as content aggregators, recycled-advice sites and businesses that monetized traffic without building a brand.
Those declines get told as “SEO is dead”, but in reality they are systems that stopped working when AI made thin, interchangeable content redundant.
The problem with alarmism is that it pushes CMOs toward the wrong reaction. Organic investments are slow, cumulative and hard to defend in moments of panic, so they’re often the first to be cut. It may seem rational in the short term, but it’s the exact opposite of what’s needed when brand authority matters more, not less.
In an AI-driven landscape, organic visibility is no longer just about clicks. It serves to build credibility, consistency and presence in the broader ecosystem the models draw from. When companies cut investment in content, technical foundations and long-term visibility, they aren’t protecting themselves from change: they’re writing themselves out of the narrative.
The sites suffering most today aren’t the ones building expertise and trust, but the ones that never did. Brands that keep investing in organic are building lasting authority. Those that pull back may not notice right away, but they’ll feel the effect when they discover they no longer control how they’re represented.
With the introduction of advertising in ChatGPT, answers stop being a “neutral” space and become commercial territory. Isn’t there a risk that brands will fool themselves into thinking they’ve solved the visibility problem by paying to appear, when in reality they’re just renting attention without building lasting authority or trust in the eyes of algorithms and users?
The risk exists, but we need to be precise about what we know and what we don’t yet know.
Using paid visibility while building organic authority makes sense. A two-track approach has always had its logic: you buy attention while you earn trust. The problem isn’t the existence of ads, but the expectations about what they produce inside AI systems.
Right now there’s no clear evidence that paid presence in conversational interfaces contributes meaningfully to the long-term understanding of a brand once the spending stops. Ads create exposure in the moment, but it remains an open question whether they help the models “learn” who a brand is in the same way consistent organic signals do.
That uncertainty weighs even more heavily if costs approach those of premium media. At that point you’re not buying efficient discovery, but very expensive moments of visibility. If that visibility doesn’t accumulate and doesn’t persist beyond the placement, you risk mistaking a temporary presence for real brand growth.
Advertising can play a role, especially for awareness. But until evidence emerges that it contributes to a lasting understanding of the brand within AI systems, it should be treated as a complement to organic authority, not a substitute for it. The slow work of clarity, consistency and credibility is still what survives when the budget runs out.
AI tends to favor coherent, detailed narratives even when they aren’t true, drawing on third-party sources to fill in the gaps. A brand’s reputation is no longer built or damaged only by users, but is literally being rewritten by machines. How can a company defend itself against this kind of hallucination?
It’s one of the most important questions in this entire interview, and the answer is uncomfortable because there’s no quick fix.
The answer is this: you defend yourself by leaving fewer gaps to fill.
AI systems don’t hallucinate out of malice, but out of uncertainty. When information about a brand is incomplete, inconsistent or scattered, the model does what it was designed to do: it builds a coherent narrative using the available signals, including third-party sources that may be outdated, wrong or only vaguely relevant.
The defense isn’t about correcting the model after the fact. By the time a false narrative becomes visible, it has often already been reinforced elsewhere. The real work happens beforehand. Brands need to be explicit, repetitive and deliberately consistent about who they are, what they do and what they’re credible for. That information needs to live in the places machines actually look: owned content, documentation, authoritative profiles and reliable third-party references.
This is where human judgment becomes central again. Someone has to take responsibility for maintaining the narrative across every channel, not just publish and move on. If the website is vague, if product descriptions change tone every quarter, if external mentions are ignored, the machine will gladly step in to “complete the story”.
Defending against hallucinations isn’t about trying to suppress them, nor about applying short-term optimization tactics. It’s narrative hygiene. Saying the same true things, clearly and consistently, in enough credible contexts that the model doesn’t have to guess. Because if a brand doesn’t define itself precisely, the system will, without asking permission.
Below you’ll find the video where I comment on the interview with Carolyn Shelby and explain why it can be so useful for you:
Why brand consistency matters as much as (and more than) technique
If there’s one thread running through this entire interview, it’s that SEO is not made of tricks, nor of chasing the latest interface change, but of clarity, consistency and accountability.
We talked about the risk of losing authority, about AI used as a substitute for human judgment, about inconsistent messaging that weakens brands, and about an increasingly concrete risk: letting machines rewrite your company’s reputation (with all the problems that entails!).
In this scenario, visibility and traffic are no longer enough. What counts is what remains, what gets repeated correctly, what makes a brand recognizable and trustworthy over time. It’s a shift in perspective that concerns anyone working with content, business and digital strategy.
My impression is that this interview hits the mark because it forces you to slow down and get organized, precisely when everything around you is pushing in the opposite direction. While the public conversation keeps framing AI as a lever to scale faster, here the point becomes something else: discipline, control and identity.
It’s ground many brands prefer not to be measured on, because it demands clear-cut choices and ongoing accountability.
If you work on your business’s visibility, these answers may feel uncomfortable, because they challenge established habits and reassuring metrics. That’s exactly why they’re useful. And today, in an ecosystem dominated by speed and impulsiveness, usefulness is worth far more than reassurance.
Thanks to Carolyn for the pleasant and productive conversation, and thank you for following us all the way here. See you in the next episode of SEO Confidential, with another interview that will challenge a few more of your certainties. Don’t miss it.
#avantitutta
