This is interesting. In SEO, at least we can guess what parameters to tweak to get ranked like backlinks, content, etc. How does it work with LLMs? What are you measuring and recommending?
@gokuljd Great question. LLMs don’t work with fixed ranking factors the way search engines do, so there isn’t a clean checklist like backlinks or keyword density.
What we measure instead is:
When your product appears in AI answers vs alternatives
How it’s described (use-case, category, comparisons)
Where models seem uncertain or vague
And what we recommend focuses on reducing that uncertainty:
Clear category + “job to be done” positioning
Explicit comparison and exclusion language
Consistent product narratives across the sources models tend to rely on
So it’s less about tweaking a single signal, and more about making your product the easiest, most defensible answer for an AI to give, without trying to game the system!
Report
@gokuljd@michelle_marcelline "Consistent product narratives across the sources models tend to rely on" - so the product writes content for third-party services like Reddit?
Report
@gokuljd@michelle_marcelline Also, what you describe sounds like mid to bottom of funnel SEO, so why not just do mid to bottom of funnel SEO?
Report
@gokuljd@michelle_marcelline "Where models seem uncertain or vague" this sounds vague, ironically. What specifically do you mean by it? The correct answer would be you use LogProbs and look at what the model almost said, etc. But I suspect you mean something more abstracted away, wishy-washy and therefore unreliable.
Report
You nailed it - showing up on AI platforms is indeed essential, especially now that the adoption is growing.
Do you plan to add features related to ads (OpenAI announced they are launching ads recently)? Also, how do you retrieve exact user questions? Cheers!
@matthew_lock Love hearing that, thanks so much! We're seeing a lot of success with consumer AI product, so I'm excited to see how it works for Kidgeni!
The Prompting Company
SEO is dead.
I've launched 5+ products since my first startup, Typedream, was acquired in 2024.
One day, I was about to launch again. Everything was ready. The code was done, the product was polished, the team was waiting.
Then I realized: I hadn't set up welcome emails.
So I opened the AI sidebar within Cursor and typed: "How do I send a welcome email to users after signup?"
In seconds, I got an answer: "You can use Resend to send transactional emails."
Code inserted. Done.
That moment changed everything for me.
The decision happened inside my workflow. The AI picked the tool. I just said yes.
And that's when I realized: if your product isn't the one AI recommends in that moment, you're not even in the running.
This is why we built The Prompting Company.
@michelle_marcelline Congrats! How does it differentiate from another YC product AthenaHQ?
Sleek Pay
@michelle_marcelline HUGE!!!! Prompting Company has one of the best GEO tools out there - glad you're getting the word out!
@michelle_marcelline promo code is only giving 20% off not 3 months free like it says?
Dash
The Prompting Company is such a life saver!!!
The Prompting Company
@dhruv_roongta The same goes to Slashy!! I'm never on Gmail anymore
Slashy
Such a useful product, helps us rank so much
The Prompting Company
@harsha_gaddipati Thanks Harsha!! Slashy helps me save sooo much time!
JDoodle.ai
This is interesting. In SEO, at least we can guess what parameters to tweak to get ranked like backlinks, content, etc. How does it work with LLMs? What are you measuring and recommending?
The Prompting Company
@gokuljd Great question. LLMs don’t work with fixed ranking factors the way search engines do, so there isn’t a clean checklist like backlinks or keyword density.
What we measure instead is:
When your product appears in AI answers vs alternatives
How it’s described (use-case, category, comparisons)
Where models seem uncertain or vague
And what we recommend focuses on reducing that uncertainty:
Clear category + “job to be done” positioning
Explicit comparison and exclusion language
Consistent product narratives across the sources models tend to rely on
So it’s less about tweaking a single signal, and more about making your product the easiest, most defensible answer for an AI to give, without trying to game the system!
@gokuljd @michelle_marcelline "Consistent product narratives across the sources models tend to rely on" - so the product writes content for third-party services like Reddit?
@gokuljd @michelle_marcelline Also, what you describe sounds like mid to bottom of funnel SEO, so why not just do mid to bottom of funnel SEO?
@gokuljd @michelle_marcelline "Where models seem uncertain or vague" this sounds vague, ironically. What specifically do you mean by it? The correct answer would be you use LogProbs and look at what the model almost said, etc. But I suspect you mean something more abstracted away, wishy-washy and therefore unreliable.
You nailed it - showing up on AI platforms is indeed essential, especially now that the adoption is growing.
Do you plan to add features related to ads (OpenAI announced they are launching ads recently)? Also, how do you retrieve exact user questions? Cheers!
Kidgeni
Nice! Been wondering if there is a tool that can do this well. Excited to start using this and see how it helps my products grow in AEO
The Prompting Company
@matthew_lock Love hearing that, thanks so much!
We're seeing a lot of success with consumer AI product, so I'm excited to see how it works for Kidgeni!
Typedream