How to Choose Which LLMs You Should Track Prompts On
If you're starting with answer engine optimization (AEO), generative engine optimization (GEO), or AI visibility, don't track every platform by default. Start with your ideal customer profile (ICP), identify where those buyers research products, and match that evidence with what we know about major AI platforms.
Start With Your ICP
If you manage AI visibility for a technical B2B SaaS company, Claude, ChatGPT, Gemini, and Perplexity are useful places to test. Use your buyer data to decide what to track.

Collect Buyer Data Before Choosing Platforms
Use website analytics, sales and customer success conversations, and market research to understand where your buyers research products and what they ask. Then compare those signals with public research on AI platform usage to interpret the evidence and prioritize platforms effectively.
-
Sales data: Look for mentions of AI assistants in discovery calls and CRM notes. Record which platform prospects used, what they asked, and whether it influenced discovery or comparison.
-
Website data: Review AI referral sources, landing pages, geography, language, and audience interests in your analytics. Referral data shows where visits come from, not every place your brand appeared.
-
Market research: Interview or survey target buyers. Ask which assistants they use to discover and compare products, not just whether they have tried a tool.
-
Customer success tickets: Look for recurring questions customers ask support. Those questions can become prompts to test even when the ticket does not name an AI platform.
Use an LLM to group anonymized notes into themes, then check the output against the original calls, tickets, and analytics. You want a short list of buyer tasks, candidate platforms, and evidence for each.
The table below summarizes the strongest public evidence on typical use and geography.
| Platform | Typical use/audience | Geography |
|---|---|---|
| ChatGPT | Broad mainstream use | Global |
| Gemini | Broad mainstream use | Global; 230+ countries |
| Perplexity | Research, learning, productivity | Worldwide |
| DeepSeek | Developer and open-source use | China, Russia, Iran, Cuba, Belarus, parts of Africa |
| Claude | Work and coding | US, India, Japan, UK, South Korea |
| Grok | More common among men (US) | International |
ChatGPT
ChatGPT is the world's most-used consumer LLM, followed by Gemini. In Pew's representative US adult survey, 44% reported using ChatGPT. A separate small marketing survey found 85% of respondents included it in their tool mix.
OpenAI's consumer study finds about 70% of ChatGPT use outside work, with practical guidance, information seeking, and writing accounting for roughly three-quarters of conversations.
In a self-selected web developer survey, 88.4% had also used ChatGPT, making it a strong baseline for both general and technical prompts. (OpenAI usage study; State of AI survey).
Gemini + AI Overviews
Pew found 24% of US adults reported using Gemini. In separate self-selected surveys, 72.6% of web developers had used it, and 44% of marketing respondents included it in their tool mix.
AI Overviews is a feature within Google Search, available in more than 200 countries and territories, so track it for relevant search queries rather than treating it as a separate chatbot audience, where search-based discovery occurs, versus conversational AI Overviews.
Claude
Anthropic's Claude.ai analysis finds work the most common broad purpose and coding a prominent task. The US, India, Japan, UK, and South Korea lead overall usage.
In a self-selected web developer survey, 82.1% said they had used Claude. This supports tracking technical and developer prompts. (Anthropic Economic Index; State of AI survey)
Pew found 6% of US adults reported using Claude, while 29% of respondents in a small marketing survey included it in their tool mix. (Pew US adults; Brafton marketers)
Grok
A Pew survey found that 8% of US adults reported using Grok, with use at 11% among men and 4% among women. (Pew Grok survey; xAI Series E).
Pew also reported Grok use at 9% among US adults aged 18-29, 10% among 30-49, 7% among 50-64, and 3% among 65+.
Perplexity
Perplexity's published Comet assistant study found more than half of activity was personal, with 57% of interactions involving productivity, learning, or research.
In the same self-selected web developer survey, 31% had used Perplexity; a small survey of marketing teams found 24% included it in their tool mix. (State of AI developers; Brafton marketers)
DeepSeek
A self-selected web developer survey found 30.9% had used DeepSeek. Microsoft (State of AI developers; Microsoft AI Diffusion) estimates its share of measured AI usage at:
-
89% in China
-
56% in Belarus
-
49% in Cuba
-
43% in Russia.
In a study of medical students in Anhui, China, 81.6% reported having used DeepSeek.
Start with one broad platform, then add one or two specialist platforms based on your budget and buyer evidence.
ChatGPT and Gemini are the broadest starting points. Track one or both when your buyers use them, and treat AI Overviews as a separate Google Search surface.
Add a specialist platform when buyer evidence supports it: Claude for technical or developer questions, Perplexity for research-heavy tasks, Grok where buyers use it for category research, or DeepSeek in markets where local usage is strong.
Examples of Platform Choices
A Local Grocery Business With a Small Budget
A local grocery business with a small budget can start with ChatGPT for local discovery questions and AI Overviews for relevant Google searches.
Add Gemini or Perplexity only if customer interviews, referral data, or pilot prompts show useful discovery there. Location matters more than a global usage percentage.
A European Direct-to-Consumer Skincare Brand
A Europe-wide skincare brand can test ChatGPT and Gemini for comparison questions in its key languages, AI Overviews for search queries in target countries, and Perplexity for research-heavy questions. Add Claude or DeepSeek only when buyer or market evidence supports them, ensuring your tracking aligns with regional and persona-specific behaviors.
B2B SaaS
A global developer-focused B2B SaaS company can start with ChatGPT and Claude for technical evaluation prompts, then test Gemini and Perplexity.
Measure AI Overviews separately when search discovery matters. Add more platforms only when the extra coverage justifies the cost.
![]()
Start Tracking Your Prompts
Once you've chosen the platforms, add a prompt set you can compare. Our earlier article explains how to enrich prompts with owned data, Google Search Console queries, and trend research.
OpenLens supports tracking across seven AI platforms:
-
ChatGPT
-
Gemini
-
AI Overviews
-
Claude
-
DeepSeek
-
Grok
-
Perplexity
Take control of what LLMs say about you. Use OpenLens for free and enjoy research-grade GEO visibility software.