75% of ChatGPT users rely on ‘keywords’ for local services

There is a common perception throughout the SEO industry that people are no longer searching for local services as much as they once did.
As tools like ChatGPT become more common starting points, that speculation often takes the form of keyword searches that offer longer, more conversational instructions.
To test it, we observed everyday users as they use ChatGPT to find local service providers, including health care and beauty procedures.
Participants were asked to start their search on ChatGPT and behave as they normally would, whether that meant visiting websites, checking social profiles, or reading reviews.
That observation was guided by a set of core questions:
- Are customers using ChatGPT as much as we think when they need local services?
- Are ChatGPT users abandoning keyword-style searches, rendering traditional keyword research and tracking useless?
- Do people have extended conversations with ChatGPT if their goal is work?
The following findings challenge many of the claims that are currently being repeated in conference rooms and across SEO-focused social media.
75% of sessions include ‘keyword searches’

A common assumption is that humans no longer search using keywords and instead rely on conversational intelligence when interacting with AI.
That thinking influenced our initial approach. When we started looking at user behavior on ChatGPT, keyword usage wasn’t a focus at first.
In a separate user behavior study we did in Google’s AI mode, we were surprised at how many searches were similar to the keywords we tracked over the decades.
That earlier discovery led us to explore whether similar behavior appeared in ChatGPT. It does.
Of all the sessions viewed, 75% entered at least one piece of information that could be classified as keyword-based.
These screenshots from a session recording show this behavior:






We shouldn’t be so surprised.
It takes a lot less effort to type “dentist 11214” than “5 best dentist online recommendations near india street, brooklyn, new york” (SIC).
(Both are case studies. You can view the full list here.)
This is also how people have searched since the early days of search engines.
Old habits die hard – and “good plastic surgeons in the brooklyn 11214 area” give the user what they need.
These findings should prompt a re-visit of the discussion surrounding keyword tracking in GEO.
Another belief is that GEO should include a step where a commercial keyword is entered into the tool and converted into a long phrase.
For local services, this is not necessary.
While it’s true that LLM answers vary from search to search and include personalization, people still enter keywords when looking for services.
This suggests that keyword research and tracking still has a place in GEO.
Dig deep: Why local SEO is thriving in the AI-first search era
Get the newsletter search marketers rely on.
The place is not that conversation: 45% of the session was one message


Extensive back and forth with ChatGPT is not uncommon when users search for local services.
About half of these sessions – 45% – did not include follow-up questions or anything resembling a conversation.
Additionally, 34% of the secondary queries were simple requests for further results.
When searching for local resources, the average ChatGPT user employed by 2.1 is prompted.
| Work | Average number of orders |
| Find a new dentist | 2.41 |
| Find a place to get botox | 1.96 |
| Get a dermatologist to check for moles | 1.71 |
| Hair transplant | 1.33 |
| Find a chiropractor | 2.33 |
| Decided to get a fix | 2.00 |
This does not mean that there are no discussions taking place. Just saying “the use of ChatGPT chat in local search” is undoubtedly reinforced.
Dive deep: GEO x local SEO: What it means for the future of acquisition
Notification behavior varies by purpose – and can change
Undoubtedly, it appears that people have conversations with ChatGPT that do not reflect the findings of this study, but those conversations are generally informational in nature.
It is also worth noting that finding data to support this was difficult. It seems to be a generally accepted view supported mainly by anecdotal evidence.
When these users need services, they often value the model in the same way they search on Google.
This behavior can change over time.
Many of the participants in this study did not have access to the paid version of ChatGPT. Their use may change as they are used to using it.
It may also be easier to search by keywords.
If a model can provide the results users need without requiring long sentences, they will likely always choose the low-effort option.
View the full study and data for more information. (Disclosure: I’m the founder of Sagapixel.)
Dive deep: How to improve local SEO: 4 AI-driven tactics that get results
Contributing writers are invited to create content for Search Engine Land and are selected for their expertise and contribution to the search community. Our contributors work under the supervision of editorial staff and contributions are assessed for quality and relevance to our students. Search Engine Land is owned by Semrush. The contributor has not been asked to speak directly or indirectly about Semrush. The opinions they express are their own.



