Two AI Visibility Startups Are Debating Measurement. I’m More Interested in What They’re Not Measuring.
A few weeks ago, a debate between two AI visibility startups caught the attention of marketers.
It wasn’t about whether SEO is dead. It wasn’t about which AI model is better. Surprisingly, the discussion was about something much simpler: how AI Visibility should be measured.
On one side was Evertune. The company argued that if you only ask an AI the same question once a day, the results can change enough to make your visibility score unreliable. Imagine your score goes up by ten points this month.
That sounds like progress, but what if the margin of error is even bigger than that? Did your marketing actually work, or did you simply get a different answer from the AI?
Profound disagreed, but not because they thought statistics didn’t matter. Their argument was that marketers shouldn’t judge AI Visibility using only one prompt. Instead, they measure hundreds or even thousands of prompts across an entire category. When all those prompts are combined, the random changes become much smaller.
In a published experiment, the difference between running prompts once a day and ten times a day was only 0.25 percentage points.
At first, it looked like another disagreement between competitors.
But after reading both sides, I realized something.
They were both trying to answer the same question:
“How do we measure AI Visibility more accurately?”
It’s an important question.
But I couldn’t stop wondering about another one.
“Are we measuring the right thing?”
Visibility Tells You What Happened
Imagine you own a coffee shop.
At the end of the day, you count 200 customers walking through the door.
That’s useful information.
You know people are coming.
But if someone asks,
“Why are people choosing your coffee shop?”
The number can’t answer that.
Maybe people love your coffee.
Maybe a food influencer recommended you.
Maybe you’re next to a busy train station.
Maybe your customer service is excellent.
The number tells you what happened. It doesn’t explain why it happened.
I think AI Visibility works the same way.
A Visibility Score tells you how often your brand appears in AI answers. That’s valuable.
But it doesn’t automatically explain why AI chose your brand instead of someone else’s.
The Question That Changed My Perspective
While researching this topic, I stumbled upon a research paper introducing a framework called Authority & Visibility Optimization (AVO).
What caught my attention wasn’t another measurement method.
It was the question the paper asked.
Instead of asking:
“How do we measure AI Visibility?”
the paper asks:
“Why does AI recommend one brand instead of another?”
At first, those questions sound almost identical.
They’re not.
One is trying to measure the result.
The other is trying to understand the cause.
That small difference completely changed how I think about AI search.
Imagine You’re Looking for a Dentist
Let’s say you just moved to a new city.
You search:
“Best dentist near me.”
You find two clinics.
Both look professional.
Both have modern websites.
Which one would you choose?
Most people wouldn’t pick randomly.
They’d probably look for clues.
For example:
- Does the clinic have good reviews?
- Has it been mentioned by trusted websites?
- Do the dentists have experience?
- Is the business information consistent everywhere?
- Are other people recommending it?
Without realizing it, you’re looking for signals of trust.
According to the AVO research paper, AI works in a similar way.
Before recommending a brand, AI doesn’t rely on one single signal. It gathers information from many different places across the web to decide whether a brand is trustworthy enough to include in its answer.
Some of those signals include:
- Helpful and well-written content.
- Mentions from trusted publications.
- Reviews and public reputation.
- Consistent business information.
- Backlinks and references from credible websites.
None of these signals guarantee a recommendation.
But together, they help AI build confidence in a brand.
The research paper refers to this collection of trust signals as Authority.
Authority and Visibility Are Different Things
This was probably the biggest takeaway for me.
Many marketers talk about AI Visibility as if it’s the whole story.
The AVO paper suggests it’s only part of the picture.
Think about it this way.
Visibility answers questions like:
- Does AI mention my brand?
- How often does my brand appear?
- Which competitors appear more often?
Those are useful questions.
But Authority answers different ones.
- Why would AI trust my brand?
- What signals support my credibility?
- Is there enough evidence across the web for AI to recommend me confidently?
One measures the outcome.
The other helps explain the reason behind the outcome.
That’s an important difference.
The interesting part is that these ideas are no longer just theoretical.
Today, marketers can already measure both sides. Platforms like AVO by Avonetiq allow brands to evaluate not only how visible they are across AI-powered search and recommendation platforms, but also the underlying signals that contribute to that visibility through metrics such as Authority Score and Visibility Score.
Rather than looking at AI mentions alone, this makes it easier to separate two different questions:
- How visible is my brand today?
- What signals are helping—or preventing—AI from recommending my brand?
Where SEO, GEO, and AEO Fit In
Reading the paper also changed the way I think about all the new marketing terms we hear today.
SEO. AEO. GEO.
They often sound like separate strategies.
But maybe they’re not.
Instead, they can be seen as different ways of strengthening either Authority, Visibility, or both.
For example:
- SEO helps search engines discover and understand your content.
- AEO helps your content answer questions directly.
- GEO focuses on making your content easier for generative AI to use.
Each one improves a different part of your digital presence.
The AVO paper doesn’t argue that these approaches are wrong. Instead, it places them under a broader goal: helping a brand become both easier to find and more trustworthy to recommend.
That perspective made a lot of sense to me.
Instead of asking whether SEO is better than GEO, or whether GEO replaces AEO, it encourages marketers to ask a simpler question:
“Which signals am I strengthening?”
Maybe We’re Asking the Wrong Question
The debate between Profound and Evertune is valuable.
As AI becomes a bigger part of how people discover brands, marketers need better ways to measure visibility.
Better measurement benefits everyone. But if you’re interested in understanding what actually influences those measurements, the original AVO research paper is worth reading.
But after reading the debate—and then discovering the AVO research paper—I came away thinking about something else.
Maybe visibility isn’t the starting point.
Maybe it’s the result.
Before asking,
“How can I improve my AI Visibility Score?”
it might be worth asking,
“What would make AI trust my brand enough to recommend it?”
They’re different questions.
And I think the second one may become even more important as AI continues to change how people search, compare, and choose brands.
References
- Jennifer Zou & Josh Blyskal, Is Once a Day Enough? (Profound)
- Brian Stempeck, Profound Has a Data Problem (LinkedIn)
- Authority & Visibility Optimization (AVO) Research Paper
