Everyone’s talking about AI search optimization right now. Thanks to incredibly sophisticated AI tools at our fingertips, anyone can sound like an AI Search/GEO/AEO expert overnight. This means there are many opinions and generalized “best practices” being touted as THE way.
But just like the tools themselves, information presented confidently doesn’t always translate to reality.
Or, it might be accurate in some cases, but completely wrong for your industry or niche. Following generalized broad advice can take you backward or in the wrong direction, which leads to frustration that “GEO/AEO/AIO whatever you want to call it” doesn’t work or translate to business impact.
In that vein, I thought I’d share 3 common themes and findings from my 4 most recent AI Visibility Audits across very different clients:
- a private equity firm
- higher ed (a college)
- a healthcare nonprofit
- a B2B professional services company (AI recruiters)
Distinct industries and audiences. But the patterns were consistent, and a few of them push back claims currently being repeated as fact in the GEO/AEO space.
3 Takeaways From My Recent AI Visibility Audits
1. SEO Performance Results Mirrored LLM Perfomance Results
Before I jump into the AI side of things, I pull the website’s current SEO performance. I want to know what I’m walking into and give context to the data I’m about to see.
I’m also looking to validate something specific: how traditional SEO performance correlates with AI visibility, and where that correlation actually shows up.
There is some information out there that traditional SEO performance has zero correlation with how well you show up in AI answers.
My findings did show a correlation in these last few audits: companies and organizations with better SEO showed up more in AI answers.
Now again, is that the case across the board? Possibly and probably not. It’s still too early in the AI Search game for definitives, in my opinion. But SEO data helps frame the picture before I start testing prompts and head to an AI auditing tool.
And just a reminder, a strong traditional SEO performance means there is a healthy mix of branded and (meaningful) non-branded search terms. The foundation of great SEO is the same as the foundation for great GEO/AEO.
2. “Reddit and YouTube Are the Top Citation Sources” – yeah, no.
There is frequently cited research stating that YouTube and Reddit are the top two sources cited in AI responses.
While it may not be totally wrong, it’s being applied too broadly.
Guess how many times Reddit and YouTube were cited in total across all 4 audits?
You guessed it. 0.
What I primarily saw were links to website pages and posts, industry aggregators, and well-known niche industry websites (literally, niche.com for the college) with profiles and listsicles.
LLMs are context-dependent. AI isn’t just looking for popular sites; it’s looking for authoritative sites within your specific vertical as well as corroboration. If you’re a private equity firm, a Reddit thread holds less weight than a Bloomberg mention or a specialized industry database.
What shows up in your space depends on your space. So before you go all in on Reddit or YouTube because of a generalized trend, make sure those will be effective for you (of course those are GREAT strategies, just do them in the right way for the right reasons!)
3. AI Won’t Fill in the Gaps (aka the Inference Fallacy)
This was one of the most consistent findings across all 4 audits.
Most people assume AI is “smart” enough to connect the dots. It isn’t. It’s a pattern matcher. If the pattern (Your Brand = Specific Solution) isn’t explicit, the pattern doesn’t exist.
You have to decide, clearly communicate, and reinforce your brand story.
One client showed up consistently, but mostly through third-party mentions. Another performed well when named directly, but disappeared in unbranded prompts.
The issue wasn’t capability. It was explicitly stating it in clear language reinforced on the site across various places and then corroborated offsite.
Decide what you want to be known for in the language your buyers actually use.
Own it on your website, clearly and directly. AI won’t infer what’s implied or buried or invisible.
Amplify it across the rest of your presence. Social, PR, industry sites, and third-party listings (aka, off-page SEO and digital PR). This is not revolutionary or new. In traditional search, backlinks and mentions were critical to growing and maintaining visibility and rankings, partly due to EEAT (experience, expertise, authoritativeness, and trustworthiness) and other signals baked into the algorithm. In AI, these same third-party mentions help determine whether you’re included at all.
What a Short, To-The-Point AI Visibility Snapshot Actually Tells You
A simple but effective AI Visibility audit is not meant to be exhaustive or slick. There are plenty of free and paid tools that will give you a polished PDF full of scores and bar charts.
These audits hone in on decision-stage prompts that drive business impact; meaning, being visible in these prompts and topics will drive very qualified leads and sales to a site (less “What is AI recruiting” and more “Best AI recruiting firms to fill specialized data center openings at a mid-sized company”). They are meant to be fast, directional, and useful, and look at:
- Your current organic search footprint to see how it translates to AI Visibility
- The decision-stage prompts your audience is using
- How you show up across LLMs your audience uses (pick 2-3)
- What sources are being cited
- Where the most important gaps are
At the end of the day, nothing is absolute. You have to take what you hear about how to be visible in LLMs with a grain of salt and not as gospel.
Interested in seeing how one of my custom AI Visibility Audits could lead to business growth from qualified AI-search driven traffic? Reach out to me today!
GEO/SEO |