Ep 5.2: What Happens to Keyword Research in the Wake of GEO?

HomeGeek SpeakPodcastEp 5.2: What Happens to Keywor...

S05e02

Transcript

[Genny]: As much as we like to say that SEO isn’t dead, one thing that has definitely evolved in the wake of GEO is keyword research. And for those of you who’ve accidentally stumbled upon this podcast for the first time, GEO, or generative engine optimisation, is optimising your content for AI-generated responses.

SEO is not dead, but there are some accommodations to keyword research we could be making for GEO that won’t throw the baby out with the bathwater. The main reason is that when it comes to searching on ChatGPT or Claude or Gemini, we don’t necessarily speak in keywords. We speak as we normally do, give or take a few weirdos.

Now before I go any further, this is not an episode where I tell you to ditch your current keyword research efforts if they’re leading to more traffic, more visibility, and more results. This is the thing that so many scare-mongering SEO influencers are trying to do right now, is fill digital marketing managers with sensationalist information about how SEO is sooo 2020 and a waste of your time and if you don’t switch, you’re going to be left in the dirt while your competitors steal all the glory and riches.

Enough of that now. We’ll probably have to keep repeating ourselves because the hype won’t go away, but if your current SEO strategy is working for you, hold onto it. Don’t replace, but proliferate.

The visibility surface has expanded, and while generative AI tools still only represent a negligible percentage of daily search queries when compared to the big kahuna, we all know that the gap will close over time. We don’t know how long that will take. Hell, based on the state of the planet right now, who knows what the future has in store for us. We also know, based on referral traffic from platforms like ChatGPT, Claude, and Gemini, that the sweet cocktail of SEO and GEO really does work.

But where does that leave keyword research? Because even you, listener, yes you, don’t query Google like you query Gemini. You talk to Google like your grandparents would talk to the family dog, as in, the fewest amount of words possible to get your point across without the baby talk: Google, difference between moose and caribou; Google, best pizza near me; Google, free IQ test. But we tend to talk to our favourite AI search engine like we’ve kidnapped a wise elder and tied them to a chair and won’t loosen their zip-ties until we decide that our curiosity and thirst for robot companionship has been quenched.

Because we use these tools differently, you might want to consider expanding your keyword research efforts and creating content that accommodates this new type of search behaviour. To wit, traditional keyword research has a new feather in its cap.

Instead of thinking of keywords as exact-match phrases, we also want to think about the intent behind the queries, more so than we were before.

Why is the person asking this question? What problems are they hoping the kidnapped AI elder will help them solve? There are a hundred ways to ask the same question, which is why GEO keyword research is less about word-for-word matching and more about satisfying the user’s curiosity or solving their problems, otherwise known as search intent.

Here’s a condensed version of an AI-inclusive SEO keyword research workflow:

Gather seed data from the usual sources plus AI query examples from whatever tool you use to expand your terms around your topic.

Then once you have your topic, you’ll want to organise by intent type, i.e., informational, commercial, and transactional. You’ll probably have 1000+ terms and topics sitting in front of you, and using a tool to help you categorise everything will obviously save you tons of time.

You then take your fancy long list and start clustering. Use AI to create semantic clusters based around meaning rather than just the words themselves.

For example, let’s say you operate a wilderness lodge and campground up near Dead Man’s Flats, Alberta.

Bearing in mind I’ve never been to Dead Man’s Flats, Alberta, here’s a shortlist of intent clusters you might capitalise on:

  • Moose sightings
  • Bear attack prevention
  • Winter accessibility
  • RV hookups
  • Good fishing spots

Now, if I’m clustering, I’d take pain point number one, moose sightings, and build my list of questions that a person may ask concerning this topic, such as:

  • How do I photograph moose ethically in the Alberta Rockies near Dead Man’s Flats?
  • What’s the best season and method for moose viewing near Canmore, Alberta, including guided tours from local lodges?
  • I’m planning a 3-day trip from Dead Man’s Flats to see moose in Banff this fall—what’s the best route, times, and safety tips?

Now of course, this isn’t an exact science because AI companies are guarded about their data, but even if they weren’t, human beings are curious and easily distracted creatures. Whatever question was asked to make you appear as a citation in an AI-generated answer might have been the tail-end of a three-hour long conversation that had practically nothing to do with anything you offer, and it’s impossible to create a totally accurate strategy that can truly accommodate the human brain’s capacity to go on tangents.

For example, how are you supposed to know that by me asking ChatGPT ‘when is the next full moon?’ I’m really trying to plan a midnight moose sighting expedition on a night with maximum visibility.

Once you’ve created your clusters, it’s time to check not your nighttime visibility, but your online visibility.

Mosey on over to ChatGPT, Gemini, and Perplexity and pop your queries in these bad boys to see what comes out. The idea being that these three platforms are built on different technologies, so avoiding Claude and Copilot also helps avoid duplication, which makes your data theoretically more accurate.

Are you already showing up for these terms? If not, which dirt bag is taking your place? What does their page have that you don’t?

You probably won’t show up in every query. Heck, you may not show up in any query. But you’ll probably show up somewhere. If you’ve been doing your SEO properly, chances are, you’re going to show up. And if you notice that you’re absent from conversations that you should definitely be in, or present in conversations but incorrectly, more work on both SEO and GEO needs to be done. But this is not an episode about backlink audits, so let’s get back to the keyword conversation.

Now that you’ve assessed where you stand and you’ve got your golden list of golden topic clusters, now you may go forth and multiply, and by multiply, I of course mean, create content.

But that golden list is probably a mile long, so to avoid overwhelm, prioritise your prompts or questions by value, by frequency, by opportunity.

  • Those high commercial intent prompts, i.e. ‘Find me an RV campground in the Alberta Rockies for May,’ are flirting at the conversion line and are pretty high value.
  • Whereas prompt clusters with high volume can lead to more visibility, albeit with stronger competition.
  • Prioritising by opportunity means addressing those obvious holes in your visibility where your competitors are already present but you’re not.

FYI, not to shamelessly plug the company that lets me talk into a microphone every few months, but if you’re not sure that you’ve got the right setup or tools to perform AI-inclusive keyword research, the folks at Geeky Tech actually have this down to a science, so get in touch with us at team@geekytech.co.uk if you need assistance setting this up. Tell them the moose lady sent you.

Now it’s time to get cracking on content. If you don’t already have pillar content, now is the time to build it. Create a core page around a main topic. Make it big and comprehensive, adding enough juicy details to give the reader real value and to signal your authority to Google and LLMs.

In our shamelessly Canadian example, that might look something like, ‘The Ultimate Guide to Moose Spotting in the Alberta Rockies’.

The purpose of this page is to demonstrate your topical authority, as I said. If you’re looking for hot tips on how to write LLM-friendly content, I will link you to our last episode where I discussed E-E-A-T, otherwise, the short version is:

  • Write conversationally
  • Answer any questions right after you ask them
  • Be as clear and readable as possible (write like Hemingway, not like Pynchon, unfortunately)
  • Back your facts with citations
  • Bring something new to the table; don’t just regurgitate what you’ve read on someone else’s page; it seems like both humans and robots appreciate value.

I know this isn’t the most utilitarian piece of advice, but since this isn’t an episode about writing AI-friendly content and I refuse to be cowed, let me just say that when you understand and embody the purpose of creating these pages, common sense will take you a long way.

Your homework this week: start with one cluster. Think of one problem or question your audience has that your company can help with.

That’s all for now. If you have any questions about how to better integrate GEO into your existing SEO strategies, you can reach us on social media at geekytechgeeks, or contact us directly at team@geekytech.co.uk.

See you next time.

Get all your digital marketing questions answered (without having to ask).
Tune in to listener-led episode.
Blog post S01E06

Show Notes

In this episode of SEO Unfiltered, we revisit one of SEO’s most fundamental tasks—keyword research—and explore how it’s evolving under the rise of GEO.

Host Genny cuts through the noise of “SEO is dead” hype to show how you can adapt traditional keyword strategies for AI-driven search engines like ChatGPT, Claude, and Gemini—without abandoning what already works.

If you’ve been wondering how to align your SEO workflows with emerging AI-driven search behaviour, this one’s for you.

Key Topics Covered

  • What GEO (Generative Engine Optimisation) actually means and how it affects search visibility.
  • Why keyword research isn’t dead—it’s just evolving.
  • How users interact differently with generative tools vs. traditional search engines.
  • The new mindset: focusing on search intent and problem-solving, not just exact-match phrases.
  • A practical workflow for AI-inclusive keyword research using a real-world example.
  • How to prioritise keywords and prompts based on value, opportunity, and visibility gaps.
  • Building pillar and supporting pages that reinforce topical authority for both SEO and GEO.
  • Why interlinking your clusters boosts crawlability, rankings, and AI citation potential.

Resources Mentioned in This Episode:

Support your fellow marketing geeks! Follow us on Twitter, Facebook, and Instagram @GeekyTechGeeks for all things SEO/GEO and advertising related—and while you’re at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

Helpful Links

Share This Post
Facebook
LinkedIn
Twitter
Email
About the Author
Picture of Genny Methot
Genny Methot
Genny Methot is Geeky Tech’s storyteller. She heads up our social media content, blog posts, and the Geek Speak podcast. Click here to learn more about Genny.
Shopping Basket