Insights · Field notes
You cannot win AI visibility with one tactic or one platform, because the engines keep changing what they trust, usually without announcing it. The only thing that survives the changes is a broad, consistent, verifiable presence, tested against the live engines. Here is a real example of the ground moving under everyone's feet, and why staying current is not a distraction from the work. It is the work.
Published July 13, 2026
"Truth is the daughter of time, not of authority."
Francis Bacon
Here is one concrete case that shows how fast the ground moves. It is about reviews, but the reviews are not the lesson. The lesson is how quickly a settled answer stopped being settled.
Reviews have become a clear factor in which businesses AI engines recommend. Ask a question like "best dentist near me" and the engines lean on review signals to build their short list. That much is settled. The interesting part is the question underneath it: which reviews, from which platforms?
Most business owners assume the answer is Google. They have spent years building Google reviews, and it is reasonable to think those carry over. But the independent research tells a more complicated story. When one local-search firm studied the listings ChatGPT surfaced, rather than asking ChatGPT to describe itself, they found the review sources skewed heavily toward Facebook and Yelp, the platforms feeding the index behind the engine, with Google Business Profile reviews historically much less visible because of how they load on the page. In their analysis, Facebook was the single most common review source on the listings that showed up.
So a business pouring everything into Google reviews, and nothing into the other platforms, may have been partly invisible to the engine on the exact signal it thought it had covered. That is not a small detail. It is the difference between being recommended and being left off the list.
Here is where the example proves the point. Just as that Facebook-and-Yelp finding was settling into accepted wisdom, it started to move. Late in 2025, people who watch this closely began reporting that ChatGPT was including Google Business Profile details, and even maps, in its local answers. There is now genuine debate in the field about whether the engines are shifting their underlying data sources, with no firm consensus. The honest position is not "Google does not matter" or "only Facebook matters." It is that the mix is moving, and anyone who tells you it is settled is selling certainty they do not have.
This is the pattern, not the exception. The engines change what they read, how they weigh it, and which sources they trust, on their own timelines, usually without an announcement. A tactic that works this quarter can quietly stop working next quarter. The businesses that stay in the answers are not the ones that found the one trick. They are the ones with a broad, consistent, verifiable presence that holds up no matter which way the engine leans.
From the founder
We tend to believe what we already think. We read the article that agrees with us, we trust the source that tells us what we want to hear, and after enough repetition, we start mistaking repetition for truth. AI engines are no different. They learn from what has been written, documented, and repeated across the web.
About eighteen months ago, working on a project in ChatGPT, it confidently gave me an answer I knew was wrong. I challenged it, explained why, and it agreed at once. So I asked what stuck with me ever since: if someone else asks the same question tomorrow, will the answer be any different? The honest answer was no, not until the sources it had learned from changed. It wasn't inventing a false reality. It was reflecting the one that had been documented often enough to trust.
Ever since, I challenge AI, not because I distrust it, but because I understand how it learns. The truth is usually out there. But truth alone isn't enough. Until it's documented, cited, and repeated by sources the engine trusts, even the smartest AI in the world may not see it.
If the ground keeps moving, the response is not to chase every shift. It is to build the things that hold up across all of them. A few principles we keep coming back to, because they survive the changes:
Build a balanced reputation, not a single-platform one. Reviews and mentions across the range of sources the engines draw from, Google, Yelp, Facebook, the directories, beat a tall pile on one platform. When the weighting shifts, a broad footprint is still standing.
Make your information machine-readable and consistent everywhere. The engines reward businesses they can confidently identify and describe, the three signals every engine looks for. Consistent details and clean structured data are the foundation that every change in ranking signals still rests on.
Test against the live engines, regularly. The only way to know whether you are in the answer is to ask the question your customer would ask and see who gets named. Not once. On a rhythm, because the answer changes as the engines re-crawl and competitors move.
Be skeptical of certainty. Anyone who promises a fixed formula for AI visibility is describing a moment, not a method. The method is staying current and building broad.
People sometimes ask why AI visibility is not a one-time fix. This is the answer. The engines are not finished products with stable rules; they are changing systems, and the sources they trust move under everyone's feet. We watch those shifts, we test against the live engines rather than trusting what they say about themselves, and we build the broad, honest, verifiable foundation that keeps a business in the answer regardless of which way the next change leans. That is the work. The day it stops changing is the day it becomes a checklist. That day is not close.
We publish what we learn doing this work, including the parts that complicate the easy story, because the honest version is more useful than the confident one. If you want to see where your business stands against the engines today, that is where it starts.
Get your CitedIQ score in about a minute. It is a free, honest read of where you stand with the six engines your customers are starting to ask, and the first step in building a presence that holds up as the rules change.
Get your CitedIQ scoreRelated reading: why cited truth wins and why ranking on Google does not mean AI recommends you. Sources: local-search review-source analysis (Whitespark, 2025) and subsequent industry reporting on AI data-source shifts (2025 to 2026). Specific figures referenced across our site are attributed on our Start Here page. The AI search landscape changes frequently; this note reflects the picture as of mid-2026 and we update our approach as it moves.
