If you work in SEO long enough, you start noticing a pattern.
Every few months, a new category of keywords appears. Recently, it has been AI prompts.
You’ll see tools claiming search volume for phrases like:
- “best ChatGPT prompts for SEO”
- “AI prompts for marketing”
- “Midjourney logo prompt”
- or even long conversational queries that look like they were copied straight out of ChatGPT
On the surface, it feels impressive. Precise numbers. Clean graphs. Trend lines going up.
But here’s the uncomfortable truth most SEO tools don’t spell out clearly:
They are not measuring how often people type prompts into ChatGPT.
They can’t.
To understand why, you first need to understand how normal keyword search volume works — and where AI prompt data fundamentally breaks that model.
Where Traditional Keyword Search Volume Comes From
For standard Google keywords, tools like Ahrefs and Semrush don’t receive raw search data directly from Google’s internal database.
Instead, they rely on a mix of:
- Google Ads Keyword Planner data
Usually range-based unless you’re spending heavily. - Clickstream data
Collected via browser extensions, ISP data partners, and anonymised panels. - SERP scraping
Observing what appears in search results and how often. - Machine-learning modelling
Used to smooth, extrapolate, and estimate demand.
That’s why even a clean number like “10,000 searches per month” is already an estimate. It’s a model, not a fact.
And that’s still the easy case.
Why AI Prompt Search Volume Is a Different Beast
When tools start showing volume for AI prompts, many people assume those numbers represent usage inside tools like ChatGPT, Claude, or Gemini.
They don’t.
They can’t — because:
- OpenAI does not release prompt usage data
- There is no “ChatGPT Keyword Planner”
- Private AI conversations are not accessible to third-party tools
No SEO tool can see what happens inside your ChatGPT session.
So what are they measuring?
What SEO Tools Are Actually Measuring
1. Google Searches
About
AI Prompts
This is the biggest source — and the most misunderstood.
When someone Googles:
“best chatgpt prompts for content writing”
That query behaves like any other Google keyword.
It:
- shows up in Keyword Planner
- appears in Ahrefs or Semrush
- gets assigned search volume
So the number reflects Google-side curiosity about prompts, not real AI usage.
In other words:
People searching for prompts, not people using prompts.
That distinction matters more than most people realise.
2. Clickstream-Based Behaviour Modelling
Some tools supplement Google data with clickstream panels.
If enough users search, click, and engage with prompt-related content, tools can model demand patterns and project volumes.
Still, this is:
- Google behaviour
- browser-observable activity
- public search intent
It is not AI-native usage.
3. Prompt-Style Query and Trend Modelling
Tools like Otterly.ai operate in a slightly different space.
They focus less on keyword volume and more on:
- AI Overviews visibility
- generative search behaviour
- domain citation frequency in AI answers
They reverse-engineer:
- prompt-style queries
- AI-triggered SERP layouts
- how often content is referenced by AI systems
But even here, they are not tracking ChatGPT prompt logs.
They’re tracking visibility, not usage.
The Three “Volumes People Confuse”
This is where most confusion happens.
There are actually three completely different things people lump together:
- Google Search Volume
How often people search a term on Google. - AI Prompt Usage Volume
How often people type something inside ChatGPT, Claude, or Gemini. - AI Visibility Frequency
How often AI systems surface, cite, or reference your content.
SEO tools mostly measure #1.
Some AI-focused tools measure #3.
Almost nobody outside the AI companies themselves can accurately measure #2.
If a tool claims it can — be sceptical.
So How
Should
You Estimate Real AI Demand?
If you’re thinking strategically, you stop chasing precise numbers and start modelling signals.
That usually means triangulating across:
- Google Trends growth curves
- Reddit and forum discussion frequency
- Prompt marketplace activity
- AI feature release spikes
- AI Overview appearance rates
- Keyword velocity rather than raw volume
You’re not measuring demand directly.
You’re inferring it.
That’s how most meaningful SEO forecasting has always worked — AI just makes it more obvious.
Why This Matters Strategically
If you’re optimising for keywords like:
- “AI SEO agency”
- “ChatGPT SEO prompts”
- “AI prompts for SEO”
Understand what you’re really ranking for:
Google-side interest in AI, not AI-native behaviour.
The bigger opportunity is not:
ranking for prompt keywords
It’s:
being cited inside AI answers
That’s a different game, with different content structures, authority signals, and optimisation priorities.
The Uncomfortable Truth
No third-party tool has:
- direct ChatGPT prompt logs
- OpenAI internal query frequency
- Gemini or Claude usage data
Anyone claiming otherwise is selling a narrative, not data.
The good news?
You don’t need perfect numbers to win.
You just need to understand what the numbers actually represent — and optimise for the right outcome.
About The Director
Notable Achievements
Struggling to Attract More Traffic?
Let us help! We provide various digital marketing services that can easily rank your website on the first page of Google such as:
Allow your website to be seen by more users by scheduling an appointment with us today!