“Visibility is like a lottery: the more your brand gets mentioned, the better your chances of appearing in LLMs and looking authoritative”
With the arrival of AI agents and agentic browsers, the moment of choice is moving further and further away from the user. Options are compared, evaluated and often decided before you can even intercept a click.
It is a profound paradigm shift: today the real challenge is not getting found, but being considered trustworthy by the model. If you work in digital, this scenario concerns you far more closely than it might seem.
In this new episode of SEO Confidential I have the pleasure of hosting Francine Monahan, Content Marketing Manager at iPullRank, one of the most authoritative names internationally when it comes to SEO, search and artificial intelligence.
Francine brings almost twenty years of experience in professional writing and content strategy, with a rare perspective: practical, clear-eyed and deeply anchored in the reality of those who have to build value, not just visibility.
With her we get into what really makes the difference today: why five-star reviews are no longer enough, what makes a brand recommendable in the eyes of an AI, why E-E-A-T still carries weight in answer engines too, and how to prevent models from filling the gaps with distorted information.
We talk about trust, omnichannel consistency, content that stands on its own, and what it means to be excluded from the decision-making process before you even reach the user.
If you have a brand active on the web, this interview is not just interesting. It is necessary.
And it will probably make you reconsider more than one certainty.
“Make it easy for LLMs to understand what you do. Don’t confuse them (more than they already are)”, Francine told us
With the advent of agentic assistants and browsers that compare options, make decisions and complete purchases on behalf of users, the discovery phase is shrinking dramatically. In this context, what concretely changes for brands that have built their growth on traffic, clicks and traditional funnels, and which ones risk being excluded from the decision-making process before they are even visible to the end user?
Your product feed and the way you write your descriptions are now even more important because LLMs can ingest and use all of them. They are the middlemen now. We need to be more descriptive and more specific in our product descriptions, explaining the value instead of sticking to the specs. And it is absolutely crucial that product information is aligned across all channels (like your YouTube videos, social posts, blog posts, etc.).
Also, generic 5-star reviews used to be fine, but now that agents are checking them, they offer no semantic value. When you request reviews, ask specific questions to guide your reviewers into leaving more scenario-based reviews that describe how they used your product and why it is useful to them.
If AI-powered answer engines synthesize information and pick a single “trusted” source, competition is no longer about visibility, but about trust. From your point of view, what factors determine whether a brand is perceived as authoritative and worth recommending by AI today?
I am still a big fan of E-E-A-T and I think AI models are too. If I am looking up some random symptom I am experiencing and spiraling into anxiety, I would rather read an article from the Mayo Clinic (an authoritative non-profit medical research organization, very well known in the US – Ed.) telling me I am not going to die than some random blog. I will be more likely to trust it. AI feels the same way.
In a recent article I wrote for iPullRank about creating an omnimedia content plan for a fictional financial company, LLMs consistently cited universities and long-standing financial institutions in their answers, so the competition for citations is fierce. You need to prove you are an expert to earn trust.
Look at what is already in the SERPs and try to create something new to stand out (like an interactive tool or a video on a topic no one has covered yet).
Structured data, FAQ and Q&A sections and proprietary data help show the models that you know your subject. Make sure your content is up to date and, if there is content on your site that is no longer relevant, delete it.
If you struggle with writing content, there are plenty of accredited freelance experts who would be happy to take on part of that work for you. Articles written by someone with a relevant professional background and a portfolio of content on the topic can definitely help demonstrate your site’s authority.
Case studies that prove your company’s success are extremely useful too, as are testimonials from happy partners/customers/collaborators, along with a broad content library covering every possible topic related to your business.
Consistency is key, though. Your message should be identical across every touchpoint, from your website and YouTube to your social media posts. AI models love getting consistent information from different places (even if they rarely provide it themselves).
In AI-powered answer engines, information about a company is often pieced together from fragmented and disparate sources. How can a brand reduce the risk of AI filling the gaps with invented or distorted data, and which official assets should it maintain to steer that synthesis in the right direction?
It all comes down to omnichannel reputation management. In my opinion it is essential that your product information, Google Business Profile, descriptions and website information are all up to date and consistent. Pay attention to reviews and to what people are saying about your business on Reddit. Respond (professionally!) to negative reviews and complaints and offer a real solution. Pretend there is a robot behind you watching your every move, because that is basically what is happening now.
We know AI does not always get it right, so it is important to make sure its training data is correct. If there are sites reporting wrong information about your brand, it is worth fighting to get it corrected.
Also, the concept of chunking content is controversial these days, but I still think organizing content in an easy-to-read format with descriptive headings is the way to go. We have seen in studies by people like Dan Petrovic that LLMs do not look at an entire page of content as a whole, but only at certain passages.
So you need to make sure that all those passages I mentioned make sense and can stand on their own.
When a user asks a “simple” question, AI search platforms break it down into dozens of sub-questions (comparisons, definitions, benchmarks, mistakes to avoid) and build the answer by drawing on different formats and channels. In your view, how do you design a content plan that truly covers that spread without producing a hundred disconnected pieces, and what “minimum” assets need to be built to become citable across the website, tools, videos and community?
If we are being technical about it, we use Qforia, a tool created by Mike King. It simulates a query fan-out and can show you a general idea of the similar queries people are searching for. It also tells you the likely search intent and offers advice on the type of content to create (a list, a video, an interactive tool, etc.). It is a great way to build a content strategy and make sure every relevant topic is covered by your content.
If we are doing it the old-fashioned way, we run a real-life query fan-out. Talk to your sales team. What do customers talk about? What questions do they ask? And do those questions lead to other questions? Look for recurring patterns and create content that can answer those questions and any follow-up questions.
Then do some research online and in the SERPs. How is it discussed in forums? What are some related topics? What about “People also ask”? Cover all the bases with well-researched, creative, interesting, high-quality content. As my manager Garrett Sussman told me, “Be so good they can’t ignore you”.
Francine, the broad query gets decided and “closed” upstream by AI Mode, even before the user sees anything. How honest, then, is the rhetoric that “if you create quality content, Google will reward you”? And what concrete signs lead us to say that a site is excluded from the initial spread (and therefore rendered invisible) not because it is worse, but because it is not compatible with the answer engine’s logic of chunks, themes and summaries?
When it comes to Google, I am a big pessimist. It makes me angry when they dictate what we should or should not do and have everyone scrambling as if they were our masters. I do not think Google ever “rewards” anyone. It rewards itself with your content and the traffic and clicks it generates. Whatever happens as a result of that is of no interest to them.
LLMs are probabilistic with their query fan-outs. As Mike King likes to say, visibility is now like a lottery. The more tickets you buy, the better your chances of winning. In the context of search, the more your brand gets mentioned (especially by a third party), the better your chances of appearing in LLMs and looking authoritative.
Bad content has always existed and always will exist on the Internet. You can tell when someone threw a blog together in a couple of hours just to have something on their site, and you can tell when someone spent days or even weeks researching, interviewing people, gathering relevant images and truly pouring their heart into a piece of content (that almost no one reads anyway).
LLMs do not quite work that way. They seem to want to see relevance. They want to see topical content in various formats that aligns with your company’s ecosystem. Could they choose to cite low-quality content? Sure. But if it really is that bad, the potential customer reading it will probably move on to another company.
Appearing in LLMs is great, but it is not everything. You also need to give people a reason to use your business beyond simply seeing it in LLMs. All in all, I do not need Google to toss a few coins onto my decomposing corpse to feel satisfied with my content.
When a brand does not get mentioned by LLMs, how much of that comes down to technical shortcomings (schema, entities, architecture, speed) and how much to a problem of identity and content consistency across the website, feeds and external channels? What signals trigger an “okay, this source is trustworthy” during the retrieval phase?
I think many of the tried-and-true technical SEO standards still hold up when it comes to LLMs. Structured data, relevant content, proprietary research, content formatted in an easy-to-read way.
When was the last time you updated your Google Business Profile? What does the description on your company’s YouTube account say? Do your LinkedIn posts sound like they were written by someone who does not even work at your company? Be consistent!
We have also seen studies showing that JavaScript can potentially prevent site content from being rendered and found by the bots that power LLMs. But even though there are some technical things you can do to improve visibility, I think content always matters a great deal.
Keeping your content fresh and up to date is essential. I do not think there is any value in publishing “mediocre” content written in a rush just because someone once said that publishing 3 articles a week is crucial. Did you write about a product or service your company no longer offers? Remove it from the site right away.
Make it easy for LLMs to understand what you do. Don’t confuse them (more than they already are).
If rankings and traffic volumes matter less, what concrete KPIs do you use to show a client that their presence in search engines is generating real value? What mistakes do companies make when they keep optimizing metrics that are easy to show but disconnected from purchase decisions?
In my opinion, the biggest mistake companies make is the one they cannot really avoid, which is catering to executives who demand to see certain metrics, even when those no longer carry any weight or meaning. It is not the CMO’s fault if the CEO keeps asking to see traffic volume every week and gets outraged when it drops, but until leadership starts listening to what their employees are saying and learning from them, we will keep seeing a lot of angry executives and stressed-out marketers.
I do not have an instant fix for this, but it might help to introduce new metrics they can get excited and obsessed about. The metrics of the future might look something like this (and you will probably need help tracking them all):
Input metrics: signals that show whether your content and your site align with how AI and search engines understand and retrieve information. These include:
- Share of Voice: your competitive visibility in the space;
- Citation frequency: how often AI search engines cite your site;
- Citation quality: whether your citations are accurate, verifiable and authoritative;
- Citation sentiment: are the mentions of your brand positive or negative?
Performance metrics: the outcomes that show whether visibility and citations translate into real business impact:
- Traffic: sessions coming from AI search and LLMs;
- Events: meaningful user actions on your site;
- Conversions: the bottom-line results;
- Engagement depth: how do users interact with your content?
With the introduction of advertising in ChatGPT, AI answers are becoming commercial real estate too. What risk do brands run when they rely solely on ads to “appear” in conversations?
A few months ago we worked on a study on user behavior in AI Mode with Farrah Bostic, who physically observed users as they searched in real time to study their patterns and preferred search methods. I watched all the recordings and wrote a series of articles about it, and one thing that struck me was how hesitant many users were to click on a sponsored link. It did not matter their age, gender, background, etc. The vast majority of participants in this study did not intentionally click on ads and openly voiced their disdain for them.
Some feared the link would take them to a scam page, others worried that their information would be stolen in a phishing attempt if they clicked. Others simply expressed verbally how annoying they find ads and how often they just scroll right past them.
I understand the importance of ads, but it will be hard to convince people that ads are safe after years of actively avoiding them. Better not to bet everything on a single card by focusing only on advertising, especially if your company is small and not very recognizable. Building a solid library of organic content remains essential, as does working on positive PR.
Some studies show that traffic from LLMs carries limited weight and converts less than organic search, others say the opposite. How do you see it? What is the most common strategic mistake you see brands make? Are ChatGPT (and the other LLMs) “new sales channels” or should we think of them more as a pre-purchase, trust-building phase?
Right now there are plenty of interesting studies showing that Google is still king and LLM traffic is fairly low, but that does not mean it will always be that way. The adoption rate of LLMs is insane. In its first month, ChatGPT had about a million weekly users. And today, more than half of adults in America use some kind of LLM.
A few months ago I read a study on Forbes claiming that conversion rates are better with LLMs than with organic search, while other studies have shown that conversion rates are roughly the same. That proves you really cannot ignore LLMs. Overall, user behavior is evolving and LLMs are constantly changing, so I do not think anyone should ignore them completely.
Things will keep changing, that much is fairly certain. The technology will improve and hallucinations will decrease (hopefully). With shopping now built in, people will end up spending more time directly inside an LLM and will have fewer reasons to search for something on Google.
I think younger generations will start using LLMs at an earlier age and AI will become part of everything we use. Not everyone will use it, of course, but I think most people will. So I believe the mistake is thinking that things will stay the same just because right now the change does not feel very dramatic.
Below you will find the video where I tell you why this interview with Francine can be genuinely useful for you and your business’s visibility in the age of AI:
With advertising on ChatGPT, does the tune change? Not a chance: it is still a matter of trust!
Have you read Francine’s words?
Artificial intelligence is not one more channel, it is a new decision-making layer.
Before the click, before the comparison, often even before the user. And that is where the most important game is being played.
Francine steers the conversation away from abstractions and brings it back to where it really counts: how a brand is understood, synthesized and judged by the models. Product feeds, descriptions, reviews, editorial content, public profiles, external mentions.
Everything contributes to building a single identity.
When that identity is consistent, rich and verifiable, AI recognizes it as trustworthy. When it is vague or fragmented, it interprets it according to logic the brand does not control.
Sure, now with advertising on ChatGPT, investing in ads will guarantee presence, but trust, I am sorry, is born elsewhere.
Users give value to recommendations they perceive as credible, not to ads. This shifts the center of gravity of any strategy: from bought visibility to built reputation.
Those who invest only in ads show up. Those who invest in authority get chosen.
On the metrics front, too, the message is very clear. Rankings and traffic volumes explain less and less. What becomes central are indicators capable of telling how, and how much, a brand makes its way into AI answers and summaries.
Share of Voice measures the real space occupied compared to competitors, citation quality indicates whether the brand is used as a trusted source, sentiment tells the tone in which it is mentioned, mention consistency shows whether its identity stays stable across sites, reviews, media and different platforms.
These metrics do not just say whether a brand is seen, but whether it is recognized, understood and considered credible. They are harder to measure, but they describe far better the real value a brand is building over time.
It seems clear to me: in the AI era, the audience rewards those who can explain well who they are, what they do and why they deserve trust. You do not need to produce more. You need to build better.
Because today, as you read these very lines, someone is already deciding on the user’s behalf.
And that someone, if you have done your work properly, will mention and recommend your brand.
Well, we have reached the end of this new episode of SEO Confidential, we thank Francine from the bottom of our hearts and see you in the next episode!
#avantitutta
