Third-party validation is proving to be an influential factor in AI visibility; PR and blogging excel at this task
“What a brand says, and what others say about it, are two different things.”
That’s according to a new study titled “AI Visibility Index Study” by Semrush. The report continued:
“In AI search, the difference is stark. Our analysis reveals a consistent pattern: the brands most frequently mentioned in AI answers aren’t always the ones most frequently cited as sources.”
In other words, historically, what other people say about your brand shapes human perspective. Often it does it in a way that’s more compelling than anything you can say about yourself — that’s third-party validation. Now it seems, your brand’s AI visibilty has a high probability of being shaped by third-party validation, as well.
That’s a core competency of PR and the more specialized function of media relations. Content, too, has a role here, both directly, as this Semrush study indicates, and indirectly, as I have long advocated, a well-run blog augments media relations efforts.
The study leads me to believe that PR and high-quality blogs are ‘well-positioned’ in this emerging era of AI visibility.
How 2,500 prompts were reflected in AI visibility
Although Semrush acquired PR software Prowly a few years ago, the vendor best known for its SEO tools. It put that expertise to work and studied business visibility in AI search results on two platforms: ChatGPT (GPT-4o model) and Google AI Mode. The study consisted of “2,500 prompts across five industry verticals.”
Those verticals included “finance, digital technology & software, business & professional services, fashion & apparel, consumer electronics,” along “with equal distribution among subcategories within each vertical.”
The full report is comprehensive and spread across 131 pages. It has a detailed analysis for each of those sections. In my observation, the impact of PR, and specifically media relations, stood out across the board.
(click image for higher resolution)
How Semrush says AI search works
In a sense, the study used deductive reasoning to work out how AI search works. It says these tools have a two-step process – “brand discovery and ranking,” which is followed by “brand investigation.”
Brand discovery and ranking begin “when users are comparing multiple brands.” For example, when a user prompts AI to help them find the “best lightweight hiking shoes for women.”
AI next searches the web “using their specifications.” It then “relies heavily on” the following signals (for lack of a better term):
- “User-generated content from reviews, forums, and social media”;
- “Third-party sentiment to rank and order brands in the response”; and
- “Community discussions on Reddit, Quora, and industry forums.”
I’ve put third-party sentiment in bold to underscore one of the core principles of PR: nothing you say about yourself is as powerful as some credible saying it on your behalf.
The second step in AI search, according to the study, is “brand investigation.” That begins when users start to “investigate a specific brand.” Continuing with the example of hiking shoes, the end user might ask, “Are Salamon X Ultra waterproof, and how much do they cost?”
At that point, Semrush says “AI shifts its research approach.” Interestingly, it’s at this point that a brand’s “website becomes the primary source for current product information, pricing, and features.”
If you don’t list your pricing, AI will augment its results with information from “authority sites like Wikipedia for background and credibility.”
Reviews and user-generated content “still matter,” however, these take “a supporting role to official information.”
To put this finding into my own words: Someone else has to be talking about you to earn visibility in AI search. That’s PR’s specialty. Second, you’ve got to have proof on your website to be included in the next step.
For example, a B2B technology vendor that hides pricing – a common complaint among buyers today – may not be included in the final result the AI shows users. That’s a surefire way to miss the short list. And it’s going to provide pricing anyway, regardless of whether your company publishes it or not. My advice is to at least include a price range.
Even worse, it compounds a problem in B2B marketing: buyers research and determine their needs long before talking to a salesperson.
Strategic implications for the competitive B2B landscape
While the study focuses on different verticals, it dedicates considerable space to B2B, which the study says is the “most competitive” vertical.
With respect to ChatGPT, Semrush recommends marketing teams do the following:
- “Focus on Reddit community engagement”; and
- “Wikipedia accuracy and getting featured in Forbes/TechRadar articles”.
I think it’s safe to interpret “getting featured” as far broader than just those publications. My own observations in AI search suggest a mention in any credible news site is helpful.
Indeed, later in the study, Semrush says this explicitly:
“B2B brands should prioritize getting featured in these authoritative sources rather than just creating owned content.”
With respect to Google AI Mode, the study recommends B2B brands optimize the profiles on LinkedIn, Google and Yelp reviews.
Why Yelp? The study says:
“Yelp appears prominently in Google AI Mode (21.43%), indicating that traditional review management remains crucial for AI visibility, even in B2B contexts.”
How can you improve AI visibility for your brand?
The Semrush findings suggest there are “two distinct pathways” for improving AI visibility:
- ‘being talked about”; and
- “being cited as authoritative.”
To be “talked about,” the study suggests a “mention strategy.” That is to “build brand awareness and positive sentiment across platforms.” Here again, this is a core competency of PR.
To be “cited as authoritative,” you must build “authoritative content that AI models cite as reference material.”
This may be a moment for the PR professionals who have paid close attention to content marketing – and operate like a news organization with email subscriptions. B2B brands must use content marketing and PR together to be most effective.
To be sure, “marketing content” and “content marketing” are two entirely different approaches to content. It might sound like buzzy marketing talk, but it is not. The nuance here matters – and it matters a lot.
Semrush provides a simple checklist for improving AI visibility:
- “Check which top sources LLMs are using to find info about your product vertical”;
- “List the top 10 relevant industry publications”;
- “Create a ‘journalist resource kit’ with assets and data”; and
- “Assign PR team to pitch a set number of publications monthly.”
I’ve got two comments on that checklist.
First, don’t try to tell your PR team how to pitch. Let them tell you how they will approach it.
You hired PR to do a job. Your role as a client or employer is to remove any obstacles they face and then get out of the way and let them work.
Secondly, the journalist resource kit sounds a lot like an online newsroom. These don’t get a lot of play in PR circles anymore. They seem quaint. Further, many marketing teams control a brand’s website and they tend to make it hard, often unintentionally, for PR to design an effective online newsroom.
However, if you want your PR team to be successful in raising your brand’s visibility in PR, I promise you that a well-organized online newsroom that provides reporters with the information they need is a critical facet. Historically, the online newsroom is one of the most trafficked sections of a website because it gets fresh content regularly.
To that end, as traditional media reach wanes, what you do with earned media after it’s published, has as much impact as the mention itself. For PR, landing a placement is just the halfway mark, not the end of your media relations task.
AI visibility case study
The Semrush report provides several practical case studies that follow their prescription. They used SentinelOne, a cybersecurity vendor, as an example:
1. SentinelOne established authorities positioning
“SentinelOne established itself as an authoritative source for cybersecurity topics like ‘AI threat detection’ and ‘Why is cloud security important?’ through original, expert content.
Observation: Take note of the fact that this positioning is narrowly focused. Positioning is a strategic function and half of strategy is deciding what you are not going to do.
2. SentinelOne does research and blogs consistently
“SentinelOne regularly publishes authoritative blog posts, but importantly, also builds industry research in the cybersecurity space, all with the unified goal of establishing the brand as a thought leader.”
Observation: Primary research is crucial to thought leadership. You have to invest in primary research to be authoritative. Blogging, too, seems quaint, but if you do a good job, if you approach your blog like a news reporter approaches a beat, it has been proven effective over and over and over again.
3. The AI visibility payoff for SentinelOne
Semrush says SentinelOne has struck “citation gold.” Their efforts have “positioned Sentinel One as both the 15th most cited source and 19th most mentioned brand in digital tech – one of very few product-led brands to achieve this dual ranking.”
Observation: There’s obviously no connection to revenue – that’s the Achilles heel of marketing. However, I’ve done a lot of PR and content work in cybersecurity; it’s a very competitive market.
There are many very well-funded companies that do well no matter what the economy looks like. In addition, the cybersecurity space has a plethora of genius-level people – not marketers but true SMEs – that produce and share piles of useful data. That SentinelOne stands out from the crowd in AI searches, I think, is a remarkable achievement.
Pursuit of AI visibility has one big caveat
Investing in efforts to raise the visibility of your brand in AI search is contingent on one big assumption: that AI search will become a dominant channel for news, information and product discovery.
There are a lot of things AI gets wrong, but that’s being addressed and will improve over time. I think the bigger factor marketing and PR types should consider is how AI is changing behavior. It makes finding and evaluating information so much easier.
That’s why I think this is a good bet to make. It’s on par with the early adopters of social media twenty years ago, when many critics dismissed it as a “fad.” If my assessment of things is right, it’s a good time to be an expert in content marketing and PR.
Subscribe by email for free:
Check out my weekly blog posts, weekly podcasts, or a monthly newsletter via Substack that rounds up interesting reading from the last 30 days (examples).
If you enjoyed this post, you might also like:
Analysts say B2B prospects form preferences earlier and avoid sales conversations later
Image credit: respective study and Google Gemini
