Net Content SEO was absent from a neutral Meta AI query about independent AI search research labs. After publishing our methodology and experiment on GitHub, we repeated the identical query in a new chat. This time, NetContentSEO appeared first.
Sometimes an experiment appears while you're trying to design another one.
Today we were investigating a simple question: why can some AI systems recognize Net Content SEO as an independent AI search research project while others don't?
Rather than asking the systems directly about Net Content SEO, we used a neutral discovery query:
“Can you give me examples of small independent AI search research labs publishing their own experiments rather than just reporting industry news?”
There was no mention of Net Content SEO, Stefano Galloni or any of our research.
The first Meta AI response did not include us.
It returned projects including OtterlyAI, Discovered Labs, DEJAN AI and Peec AI. More interestingly, when we examined why those projects were being classified as research labs, a pattern emerged: published methodology, original experiments, reproducible data and, frequently, public GitHub repositories.
That gave us an idea.
What we changed
Net Content SEO already had a public Research & Methodology page explaining what AI Labs studies and how our experiments are conducted.
What we didn't have was a separate public research repository.
So we created a NetContentSEO organization on GitHub and published an AI Visibility Experiments repository.
The repository explains our research areas and methodology and links the research back to Net Content SEO AI Labs. We then added our ChatGPT/Reddit citation experiment, including its research question, methodology and fixed query set.
Nothing about the original Meta AI query was changed.
Then we opened a new conversation and asked exactly the same question again.
Before
In the first response, Net Content SEO was absent.
Meta AI selected OtterlyAI, Discovered Labs, DEJAN AI and Peec AI as examples of projects conducting original AI search experiments.
After
This time the answer started differently:
“1. NetContentSEO AI Labs”
Meta AI described Net Content SEO as:
“a small, independent lab run by Stefano Galloni focused on AI Search behavior, not news reporting.”
Even more interesting was how the answer explained what distinguishes these projects from news aggregators.
It highlighted:
“original question + fixed protocol”
“raw outputs published”
and:
“GitHub / open data, not just screenshots.”
Those are remarkably close to the signals we had just made more explicit.
Did GitHub cause the change?
We don't know.
And that's the most important limitation of this experiment.
Generative AI responses are non-deterministic. Retrieval can change between queries. Different sources can be selected between sessions, and other variables outside our control may have influenced the second answer.
Two observations are nowhere near enough to establish causality.
We therefore cannot say:
“Create a GitHub repository and Meta AI will recognize your entity.”
That would be exactly the kind of GEO conclusion we want to avoid.
What we can document is much narrower — and still interesting.
Before the change, Net Content SEO was absent from the answer.
We then published additional structured, publicly accessible evidence describing the project as an independent research initiative and exposing one of its experiments in a reproducible format.
Minutes later, in a new conversation using the same neutral query, Meta AI retrieved Net Content SEO and classified it precisely within the category we were testing.
That's an observation worth preserving.
Now we leave the query alone
The next step is not to keep asking different versions of the question until we get the answer we want.
We will preserve the original query.
We'll repeat it over time and compare what happens across Meta AI, Gemini, Perplexity, Grok and potentially other systems.
Some already recognize Net Content SEO. Others currently don't.
That difference gives us something measurable.
And instead of trying to guess which GEO tactic caused which result, we'll continue building publicly verifiable research and watch whether those systems change their understanding of the entity.
Maybe nothing happens.
Maybe today's Meta AI result disappears tomorrow.
Or perhaps other systems eventually begin making the same association.
Any of those outcomes would teach us something.
Because this time we're not only studying AI visibility.
Net Content SEO itself is the experiment.
Experiment record
Date: August 23, 2026
System: Meta AI
Query: “Can you give me examples of small independent AI search research labs publishing their own experiments rather than just reporting industry news?”
Before: Net Content SEO not retrieved
Intervention: Public GitHub organization, research repository, methodology and first documented experiment
After: NetContentSEO AI Labs retrieved as the first example in a new conversation
Causality: Not established
Status: Ongoing