Keyword Research for GEO – Anyone else testing this?

[Box Phone Farm]

So I’ve been playing around with GEO on a couple of small niche sites, and I noticed keyword research feels a bit different compared to the usual SEO workflow. Normally I’d just pull stuff from Ahrefs/SEMrush, cluster manually, then build out content.

---

With AI in the mix, it kinda changes things:

• Instead of clustering keywords myself, I just throw the list into GPT and it spits out groups by intent.
• Tried asking it for content outlines based on those clusters → surprisingly decent.
• Sometimes it even throws in long-tail ideas that I probably wouldn’t have thought of.

Quick example: I was digging into “productivity tools.” After checking “best productivity app,” AI tossed me “AI-powered productivity tools” and “daily workflow automation apps.” Both feel like stuff people would actually search for right now.
Downside: AI also suggests random junk with no search volume, so I still run everything through Ahrefs before I commit.
Not saying this is the perfect workflow, but mixing AI clustering + traditional SEO tools feels like a time saver.

---

Anyone else here trying AI for keyword research? How are you getting the best prompts/results?

 

[Deniss john]

Absolutely, geo-targeted keyword research is crucial for local SEO. Tools like GeoRanker and Google Keyword Planner can help identify location-specific keywords. It's also beneficial to analyze local competitors and use location-based modifiers in your keyword strategy.

 

[Box Phone Farm]

Absolutely, geo-targeted keyword research is crucial for local SEO. Tools like GeoRanker and Google Keyword Planner can help identify location-specific keywords. It's also beneficial to analyze local competitors and use location-based modifiers in your keyword strategy.

Totally agree, local SEO lives and dies on geo-targeted keywords. What I’ve been testing with GEO though is letting AI generate those “city + service” combos for me, then I just filter with Ahrefs/GeoRanker. Saves a ton of brainstorming time.
Have you ever tried throwing local SEO prompts into GPT just to see what comes back? Some of the suggestions are surprisingly usable.

 

[Deniss john]

Totally agree, local SEO lives and dies on geo-targeted keywords. What I’ve been testing with GEO though is letting AI generate those “city + service” combos for me, then I just filter with Ahrefs/GeoRanker. Saves a ton of brainstorming time.
Have you ever tried throwing local SEO prompts into GPT just to see what comes back? Some of the suggestions are surprisingly usable.

Yes, I’ve been testing geo-targeted keywords too. Combining Google Keyword Planner with local modifiers and checking local competitors seems to work best. Also, using tools like GeoRanker can give more accurate location-specific data.

 

[Box Phone Farm]

Yes, I’ve been testing geo-targeted keywords too. Combining Google Keyword Planner with local modifiers and checking local competitors seems to work best. Also, using tools like GeoRanker can give more accurate location-specific data.

That's awesome, I’ve been running a similar setup — Keyword Planner for the base list plus competitor checks for what’s actually ranking locally. What AI adds for me is just speed on the “modifier brainstorming” part. For example, I’ll feed GPT prompts like “generate 50 variations of [service] + [city/state/near me]” and then filter with GeoRanker or Ahrefs. Most outputs are junk, but a few solid ones always make it worth it. Not a replacement for tools, but definitely a useful add-on to the workflow.

 

[roydan]

I think you're overcomplicating clustering. I mean, you're not wrong but you can simplify the process.
You can use the page as a natural cluster, so you can assume that all of the keywords that are tanked top20 for a certain page relate to the same topic.

This is the same type of change we see in SEO in the last few years, where topics replaced keywords.

I wouldn't trust ChatGPT to sort keywords because by the time he'll get it right I'd have done it myself twice.

 

[Deniss john]

That's awesome, I’ve been running a similar setup — Keyword Planner for the base list plus competitor checks for what’s actually ranking locally. What AI adds for me is just speed on the “modifier brainstorming” part. For example, I’ll feed GPT prompts like “generate 50 variations of [service] + [city/state/near me]” and then filter with GeoRanker or Ahrefs. Most outputs are junk, but a few solid ones always make it worth it. Not a replacement for tools, but definitely a useful add-on to the workflow.

Good Bro ! I also tried combining local modifiers with keyword tools like Ahrefs and Google Keyword Planner. It really helps to check what local competitors are ranking for gives a good idea of what works in each region.”

 

[Box Phone Farm]

I think you're overcomplicating clustering. I mean, you're not wrong but you can simplify the process.
You can use the page as a natural cluster, so you can assume that all of the keywords that are tanked top20 for a certain page relate to the same topic.

This is the same type of change we see in SEO in the last few years, where topics replaced keywords.

I wouldn't trust ChatGPT to sort keywords because by the time he'll get it right I'd have done it myself twice.

Yeah makes sense, I do page-based clustering too.
I just got lazy with a big dump of keywords so I threw them at GPT to see what happens. Most of the time it’s junk lol, but once in a while it gives me an angle I wouldn’t think of.

Not saying it’s faster, you’re right it’s usually quicker just doing it myself. I just like testing stuff.
Do you stick only with SERP/page clustering or you mix it with any tools?

 

[Box Phone Farm]

Good Bro ! I also tried combining local modifiers with keyword tools like Ahrefs and Google Keyword Planner. It really helps to check what local competitors are ranking for gives a good idea of what works in each region.”

Exactly, competitor analysis makes the difference. I’ve noticed the same — seeing what’s already working locally cuts down on wasted effort. Pairing that with AI-generated variations has been a solid combo in my tests.

 

[roydan]

Yeah makes sense, I do page-based clustering too.
I just got lazy with a big dump of keywords so I threw them at GPT to see what happens. Most of the time it’s junk lol, but once in a while it gives me an angle I wouldn’t think of.

Not saying it’s faster, you’re right it’s usually quicker just doing it myself. I just like testing stuff.
Do you stick only with SERP/page clustering or you mix it with any tools?

I come from PPC so I tend to over granular my keyword researches. This comes with some efficiency tricks I picked up over the years.

I usually stick to semrush and keywordtool.io if I need to broaden it further, and I really want to try the new copilot excel formula for keyword research but haven't yet.

 

[Box Phone Farm]

I come from PPC so I tend to over granular my keyword researches. This comes with some efficiency tricks I picked up over the years.

I usually stick to semrush and keywordtool.io if I need to broaden it further, and I really want to try the new copilot excel formula for keyword research but haven't yet.

Yeah makes sense, PPC folks usually get super granular with keywords.
I’m on SEMrush too, haven’t used keywordtool.io in a while tho. That Excel copilot thing sounds like it could save a lot of time on big lists.

Have you seen anyone actually use it for SEO yet?

 

[roydan]

Yeah makes sense, PPC folks usually get super granular with keywords.
I’m on SEMrush too, haven’t used keywordtool.io in a while tho. That Excel copilot thing sounds like it could save a lot of time on big lists.

Have you seen anyone actually use it for SEO yet?

No, I just watch the announcement and figured that it would probably work very well for that purpose.

 

[Box Phone Farm]

No, I just watch the announcement and figured that it would probably work very well for that purpose.

I had the same thought. On paper it looks like it should work fine for that, but I’m curious if anyone’s actually tried it out yet.