Study Finds Most ChatGPT Users Still Use Keywords for Local Service Searches

ChatGPT Users Still Use Keywords

Despite widespread claims that AI tools have replaced traditional keyword-style searching, new observational data suggest otherwise. When people use ChatGPT to find local service providers – especially in healthcare and aesthetics – most still rely on short, keyword-based prompts rather than long, conversational queries.

The findings challenge a growing assumption in the SEO industry: that keyword research is becoming obsolete as users shift to AI-driven discovery.

How the study was conducted

To better understand real user behavior, participants were observed while using ChatGPT to find local services such as dentists, dermatologists, Botox providers, and other medical or aesthetic specialists.

Users were instructed to begin their search in ChatGPT and behave naturally. They were free to:

  • Visit websites
  • Check social profiles
  • Read reviews
  • Ask follow-up questions (if they wanted)

The goal was to answer a few core questions:

  • Do people actually search differently in ChatGPT when looking for local services?
  • Are keyword-style searches being replaced by conversational prompts?
  • How much back-and-forth happens when intent is transactional?

The results tell a different story than what’s often repeated at conferences and across SEO social media.

Keyword-style prompts still dominate ChatGPT local searches

Across all observed sessions, 75% included at least one keyword-style prompt.

Examples included short, direct searches such as:

  • “dentist in chicago”
  • “dentists montgomery”
  • “botox by a doctor”

These inputs closely resemble the same local keywords people have used in Google search for years.

The reason is simple: typing short keywords requires less effort. When users want quick answers, they choose the fastest path — even inside an AI interface.

This mirrors earlier observations from Google’s AI Mode, where users also relied heavily on traditional keyword phrasing rather than full conversational sentences.

What this means for GEO and local SEO

There’s a growing belief that generative engine optimization (GEO) requires turning keywords into longer, natural-language prompts.

For local services, this data suggests that step is often unnecessary.

While AI responses may vary slightly between users, keywords are still how people initiate discovery, especially when they’re looking for providers nearby.

Keyword research and tracking, therefore, remain relevant — particularly for local SEO and GEO strategies.

Local searches in ChatGPT aren’t very conversational

Another common assumption is that users engage in extended conversations with ChatGPT when searching for services. The data doesn’t strongly support that idea.

  • 45% of sessions were single-prompt searches
  • 34% of the second prompts were simple “show me more results” requests

On average, users used just 2.1 prompts per session when searching for local services.

Here’s how prompt usage varied by task:

  • Finding a dentist: ~2.4 prompts
  • Finding Botox providers: ~2.0 prompts
  • Dermatologist visit: ~1.7 prompts
  • Hair transplant research: ~1.3 prompts

In other words, most users are not chatting – they’re searching.

This suggests that “conversational AI search” is far more common for informational queries than for transactional, local-intent searches.

Intent matters more than the interface

People do have long conversations with AI — but usually when they’re researching, learning, or exploring ideas.

When the goal is to find a nearby service, behavior changes. Users default to what works:

  • Short prompts
  • Location-based keywords
  • Minimal interaction

Many participants in the study were also free users of ChatGPT. Behavior may evolve as users become more comfortable with AI tools or gain access to paid features.

Still, one pattern is clear: when effort is low and results are good, users stick with keywords.

Key takeaways for marketers and agencies

This data highlights several important points for local SEO and AI search optimization:

  • Keyword-style searches are not disappearing in AI tools
  • Local intent drives simpler, faster prompting behavior
  • Extended AI conversations are overstated for transactional searches
  • Keyword research still plays a role in GEO
  • Optimizing for local discovery in AI requires clarity, not verbosity

If AI tools can deliver relevant results from short prompts, users will almost always choose that path.

The bottom line

AI is changing search – but not all user behavior is changing with it.

For local services, people still search the way they always have: with keywords, locations, and minimal friction. That reality should shape how agencies approach local SEO, GEO, and AI visibility moving forward.

Old habits don’t die easily – especially when they still work.

Leave a Reply

Your email address will not be published. Required fields are marked *