I am learning SEO recently and I have been always wondering how does GPT-5's understanding of a brand evolve over time, and can the tool track changes in brand visibility or sentiment across different versions of GPT?
@justin2025 the models by foundation labs including OpenAI often get further trained and then from time to time released as newer versions. That also means that newer versions of a model might know more about recent events than previous versions (if the recent events found their way into the training material).
That said: the main way ChatGPT answers questions that require up-to-date knowledge is by doing web searches or custom data lookups in databases that have that information.
With this tool we wanted to provide a simple way to check what the model itself knows and thinks based on its training material. We found it helpful for internal use and to be honest quite surprising in some cases.
That's why we wanted to make it available to more people.
I tried it... How does the site select its competitors? Because the website I used is for a tech hackathon (Event) and it lists competitors like Procore, Plangrid, which is out of our category.
The report is a direct test for what GPT-5 knows about a brand and its competitors (without any tool use like search). This means the report does not give an answer like ChatGPT would.
It is a test for what GPT-5 knows based on its training material. So it's a bit like using a stick to poke at a black box. For well known brands the results are kind what you would expect BUT in some cases surprisingly inaccurate.
Since for most users (unless you manually switch to another legacy model) get GPT-5 nowadays we were curious what GPT-5 knows about brands.
For smaller brands the information is not very good but that's exactly why we built it: so that we can check what the model knows when it can't use tools like search because what the model knows without any help is also affecting how it searches.
Hope that helps! I am working right now on explaining the how/why behind the report better!
findable.
Hello Makers & Creators,
This is Pete from gpt5seo.com
We built this tool, because we needed to know how GPT-5 changes rankings for our SEO clients.
So, what does GPT-5 actually “know” about your brand?
Now you can find out in seconds. Just drop in your website and you’ll get:
📊 Recognition Score => how well GPT-5 recognizes your brand
🌐 Social presence & sentiment => what tone it associates with you
🥇 Recommendation likelihood => the chance GPT-5 would suggest you in a conversation
🆚 Competitive landscape => your direct & indirect AI search competitors
✅ Analysis confidence => how sure GPT-5 is about its own answers
Let me know in the comments what surprises you about your own results.
We’ve been using GPT5SEO for our SEO clients, and tested it on big brands like Nike & Tesla,...
If you like it, an upvote means the world 🚀
Let’s make our brands findable in the age of AI search.
Pete
@peterbuch Love the Recognition Score. Can I segment by locale/model family and set alerts when competitors outrank my brand?
findable.
@peterbuch @anwarlaksir love the monitoring/alert angle. We're looking into it!
Hey Peter, congrats on the launch!
I am learning SEO recently and I have been always wondering how does GPT-5's understanding of a brand evolve over time, and can the tool track changes in brand visibility or sentiment across different versions of GPT?
findable.
@justin2025 the models by foundation labs including OpenAI often get further trained and then from time to time released as newer versions. That also means that newer versions of a model might know more about recent events than previous versions (if the recent events found their way into the training material).
That said: the main way ChatGPT answers questions that require up-to-date knowledge is by doing web searches or custom data lookups in databases that have that information.
With this tool we wanted to provide a simple way to check what the model itself knows and thinks based on its training material. We found it helpful for internal use and to be honest quite surprising in some cases.
That's why we wanted to make it available to more people.
We already built it, why not share it.
Hope that helps!
I tried it... How does the site select its competitors? Because the website I used is for a tech hackathon (Event) and it lists competitors like Procore, Plangrid, which is out of our category.
findable.
@busmark_w_nika great question!
The report is a direct test for what GPT-5 knows about a brand and its competitors (without any tool use like search). This means the report does not give an answer like ChatGPT would.
It is a test for what GPT-5 knows based on its training material. So it's a bit like using a stick to poke at a black box. For well known brands the results are kind what you would expect BUT in some cases surprisingly inaccurate.
Since for most users (unless you manually switch to another legacy model) get GPT-5 nowadays we were curious what GPT-5 knows about brands.
For smaller brands the information is not very good but that's exactly why we built it: so that we can check what the model knows when it can't use tools like search because what the model knows without any help is also affecting how it searches.
Hope that helps! I am working right now on explaining the how/why behind the report better!
@__tosh Thank you for clarification Thomas :) It is more clear now.