I recently went on Edward Strum’s podcast to talk about SEO, GEO, AEO and what brands should actually be doing if they want to show up in ChatGPT, Google AI Overviews and other AI search experiences.

The conversation started with something that has been annoying me for a while. A lot of the advice being promoted as new GEO or AEO tactics is simply old SEO repackaged with a new acronym.

Build pages around topics people search for. Put relevant keywords in URLs. Write clear page titles and meta descriptions. Add useful supporting copy. Create comparison pages. Earn mentions from other websites. None of those things are bad recommendations, but none of them suddenly became important because ChatGPT exists either.

My issue isn't that people are recommending SEO fundamentals. It's that they're often being presented without any of the history or context that comes with them. When that happens, marketers can take a tactic that has worked in search for years, overuse it because someone has labelled it “GEO,” and potentially damage the same organic visibility that helps them show up in AI search in the first place.

My view is still pretty simple: good SEO is good GEO.

That doesn't mean AI search changes nothing. There are some genuinely interesting differences, particularly around query fan-out, personalization, third-party sources, brand positioning and attribution. The challenge is separating those changes from the basic SEO tactics that have simply been given a new name.

SEO is still the foundation

One thing I keep seeing at conferences and on LinkedIn is basic SEO being presented as a breakthrough AI-search discovery. A company will analyze millions of prompts and announce that pages with relevant words in the URL are cited more often, or that content closely related to the searches an AI system performs has a better chance of being surfaced.

That data can still be useful, but an SEO shouldn't be surprised that a relevant page with clear targeting performs better than an irrelevant one.

The problem is what happens next. These findings get turned into simplified checklists for marketers who don't have much SEO experience. Put the keyword in the URL. Add an FAQ. Create more pages. Add more schema. Write a paragraph for every possible prompt.

Then somebody goes to their SEO team and asks whether they're doing all of this for GEO.

In many cases, the answer is yes. We've been doing it for SEO.

Before building an entirely separate GEO strategy, I would still want to know whether the website has the basics right. Can search engines understand the page? Does the content satisfy the intent behind the search? Does the company have proper product, category and solution pages? Is the internal linking strong? Does the website have authority? Do third-party sources support what the company says about itself?

Those fundamentals haven't disappeared. In many cases, they're more important than ever.

Big brands are still getting basic SEO wrong

One of the examples we discussed on the podcast was New Balance.

I had been comparing men's running shoe category pages in Canada and, when I checked, New Balance was outranking much larger competitors including Nike and adidas. It wasn't because New Balance had created some revolutionary AI-search strategy. In fact, parts of the page could have been much better.

What they were doing was relatively basic SEO.

The page clearly targeted men's running shoes. It included supporting content around things like how running shoes should fit, how to choose the right pair and how to care for them. It also used contextual internal links to related categories.

Nike's page, by comparison, had much less useful search-focused supporting content. A lot of what was there focused on Nike technology and product messaging rather than the broader questions somebody searching for men's running shoes might have.

adidas was somewhere in between. There was more SEO copy and plenty of internal linking, but some of the page felt like SEO requirements and brand requirements had been stitched together.

That's a problem you see constantly in enterprise SEO. I've worked with major brands, and sometimes the hardest part isn't knowing what to do. It's getting a relatively simple recommendation through brand, product, legal, development and every other stakeholder that needs to approve it.

By the time the change goes live, the original SEO recommendation can look very different.

That's also why smaller brands can still compete. If you look at the New Balance page and think the copy isn't particularly good, don't copy it. Take the same fundamentals and do them better. Write stronger copy, explain what makes your products different, answer useful buying questions and use internal links naturally.

Sometimes the opportunity isn't an advanced AI-search tactic. The biggest brands in the category just aren't doing the basics particularly well.

Query fan-out is where AI search gets more interesting

One area where AI search does introduce something marketers need to pay closer attention to is query fan-out.

When somebody asks an AI system a question, that system may perform several searches behind the scenes before it constructs the final answer. So somebody asking for the “best financial reporting software” might trigger related searches around enterprise requirements, security, customer reviews, analyst sites and specific vendors.

That changes how I think about optimization.

Instead of seeing a prompt and immediately deciding to build a page around it, I want to know what searches that prompt is generating and which sources keep appearing within those searches.

For one financial software company I was working with, I saw query fan-outs involving sites such as Gartner, G2 and Capterra. Historically, the company may have invested more heavily in one of those platforms. But if Gartner starts appearing consistently while an AI system researches the queries we care about, Gartner suddenly becomes more important.

The action might be improving the profile, generating more legitimate reviews or making sure the product positioning there is accurate.

For another company, Reddit might matter more. For another, YouTube. For another, a handful of specialist publications might dominate the results.

That's why I don't think there is one universal GEO checklist. You need to inspect what the AI system is actually doing for the searches that matter to your business and work backwards from there.

A much better process is:

Important prompt → query fan-out → influential sources → gaps → actions

rather than:

Prompt → create another blog post.

Your third-party profiles need to tell the same story

One of the easiest practical opportunities is to audit how your company is described across the web.

Imagine your website says you're an enterprise platform. Now check Gartner, G2, Capterra, LinkedIn, old press releases, affiliate sites and customer reviews. Do those sources reinforce the same positioning, or are half of them still describing you as a mid-market tool?

The copy doesn't need to be identical everywhere, but the story should make sense.

AI systems are trying to work out what your company is, who it serves and where it belongs. If the evidence is inconsistent, you're making that harder.

This is one reason I think a lot of GEO ultimately leads back to online reputation management. Before AI search, you still wanted your own site, social profiles, review platforms and authoritative third-party results to dominate branded searches. Those same sources can now contribute to how an AI system describes or recommends you.

The channels are becoming more connected.

Industry pages become more useful as answers get personalized

We also discussed industry and use-case pages, particularly for B2B companies.

A generic solution page can only speak to so many people. A company selling RFP software might have one core page for “RFP software,” but the needs of an investment manager are different from the needs of a government department or nonprofit.

That can justify dedicated pages such as RFP software for investment management, RFP software for government and RFP software for nonprofits.

The same principle applies elsewhere. An email marketing platform might have specific pages for publishers, ecommerce businesses and media companies.

I'm not suggesting companies should create hundreds of near-identical programmatic pages. If you have three or four industries that represent meaningful areas of the business, though, it makes sense to explain properly how your product serves each one.

This was already useful for SEO, but personalization makes it more important.

Ask an AI system for the best CRM and the answer might be broken down into the best option for enterprise, the best for small businesses, the best for a particular industry and the best for a particular use case. You want enough evidence on your website and elsewhere for the AI to understand which segment you belong in.

You probably don't need to track thousands of prompts

Prompt tracking has become another area where I think companies are overcomplicating things.

Take these four searches:

Technically, they're four different prompts. In practice, the commercial idea underneath them is the same: best email marketing software.

That's how I tend to approach prompt tracking. I think about themes in a similar way to keywords rather than paying to track every possible variation of the same underlying intent.

I'm also much more interested in commercially meaningful prompts.

If a B2B software company gets cited for “What is gross profit?”, that might look good in an AI visibility report, but it probably doesn't mean much for revenue.

I'd rather know whether the company appears for something like “best financial reporting software for a mid-market company” or “best RFP software for an investment management team.”

Those are the questions closer to a buying decision.

More tracked prompts don't automatically give you more useful insight.

Competitive AI visibility takes more than publishing a listicle

We spent some time talking about highly competitive searches such as “best email marketing software.”

You're unlikely to win those by publishing one article called “10 Best Email Marketing Platforms.”

You still need a strong core solution page that clearly explains what your product does, who it's for, which problems it solves and why somebody should choose it. You may also need comparison or “best software” content because those formats still appear prominently in search and AI results.

But the work doesn't stop on your website.

Who's recommending you? Do you appear in G2 reports? What are customers saying on review platforms? What does Reddit say? Are creators on YouTube talking about the product? Do affiliates include you? Are industry publications mentioning you?

At the competitive end of the market, you need a broader body of evidence supporting your positioning.

Sometimes that leads to an uncomfortable conclusion. The best SEO recommendation might not be another piece of content. It might be that people genuinely dislike something about your product and you need to fix it.

You can't SEO your way out of a bad reputation

I gave one example on the podcast from a software company where implementation had become a negative theme.

People were effectively saying that implementation was difficult or took too long.

You can try to drown that out with marketing, but a better approach is to understand why the perception exists and address it properly.

We created content that explained what implementation actually involved. One of the founders talked publicly about implementation in the industry. We produced guides around implementation timelines and made sure relevant product pages clearly communicated what customers should expect.

When I compared the sentiment I was seeing later with what I had seen in older reporting, the representation of implementation was much more positive.

That's the type of AI-search optimization I find much more interesting than “add 15 FAQs because LLMs like questions.”

If an AI system consistently surfaces a negative association with your product, ask where that association is coming from, whether it's accurate and what the company can actually do about the underlying issue.

That might involve content, PR, product marketing, customer success or improving the product itself.

AI search brings all of those areas closer together.

Repositioning your brand is harder than changing the homepage

Another area we discussed was brand repositioning.

Imagine a software company has spent 12 years being known as a mid-market platform, but it now wants to move upmarket and position itself as enterprise.

Marketing can change the homepage. Product pages can be rewritten. Old blog posts can be refreshed.

That doesn't mean ChatGPT is suddenly going to place the company alongside SAP, Salesforce, IBM or Workday.

The web contains years of evidence associating that business with the mid-market: analyst descriptions, customer reviews, old articles, press coverage, comparison pages and the company's own historic content.

You can't flip a switch and make that disappear.

There's also a risk during the transition. If you remove the old positioning from the assets you control before you've built enough evidence for the new positioning, you may temporarily weaken your association with the category you used to own without yet becoming established in the new one.

You could lose visibility before you gain it back.

The same issue can appear after mergers and acquisitions. Two companies may have completely different associations built up over many years. Once they merge, the new brand structure might make perfect sense internally, but AI systems still have a huge amount of historic information to reconcile.

Brand positioning isn't simply whatever your latest messaging document says. It's the accumulated evidence about your company across the web.

Citations aren't a business metric

This might be the biggest issue I have with the current AI-search conversation.

We celebrate citations.

A brand gets cited in ChatGPT. A YouTube video appears inside an AI Overview. A LinkedIn post gets surfaced as a source.

That's interesting, but what happened next?

During the podcast, we searched for email marketing software and found YouTube content appearing in Google's AI results. One of the interesting things was that relatively new videos with very small view counts could still appear as sources.

From a search perspective, that's fascinating.

From a business perspective, the next question is much harder.

Imagine I ask a marketing team for $20,000, $30,000 or $40,000 to sponsor creator content. Three months later, I tell the board that the video appeared as the third citation in an AI Overview.

They're going to ask how much money it made.

And attribution is extremely messy.

Someone might start in ChatGPT, perform another Google search, watch a YouTube video, read Reddit, ask another AI question, visit your website, leave, then return three days later through branded search and request a demo.

Which touchpoint gets the credit?

There isn't a perfect answer.

That's why I don't think an increase in citations by itself is enough to prove that an AI-search strategy is working.

Ask customers how they found you

One of the simplest ways I've found to get better attribution is self-reported attribution.

Add a “How did you hear about us?” field to your forms and include AI as an option.

If somebody selects AI, you can go one step further and ask what they searched or what they asked.

Now the information becomes much more useful.

Instead of knowing that ChatGPT sent 100 sessions, you might learn that a buyer asked for financial reporting software with a specific requirement, saw your company recommended and later became an opportunity.

That tells you something about the type of prompts that are actually producing customers.

It also gives you data that referral traffic alone won't capture. Somebody may discover you through an AI system and come back later through branded search or another channel.

For one company I discussed during the podcast, this type of self-reported attribution was important because direct LLM referral traffic only represented a small part of what we were seeing. Without asking customers themselves, it would have been much harder to demonstrate whether AI discovery was contributing to revenue.

There is no universal GEO checklist

The biggest takeaway from the conversation is probably that there isn't a 10-step GEO formula every company can copy.

For one company, the biggest opportunity might be improving its Gartner profile. For another, it's G2. Another company may need better industry pages. Another might discover that YouTube dominates the AI results in its category. Another might realize that customers are saying negative things on Reddit because the product genuinely has a problem.

You have to start with the business.

Then you decide what to do.

That's a strategy.

Adding FAQs because somebody told you they're good for AEO isn't.

What I'd focus on in 2026

If I were reviewing a company's AI-search strategy today, I'd start with SEO fundamentals and the parts of the website closest to revenue. Make sure important product, category and solution pages are well structured, clearly targeted and properly linked together.

From there, I'd identify the commercially meaningful query themes the business wants to influence. I wouldn't try to track every possible prompt variation. I'd focus on the types of questions a real customer might ask before making a purchase.

Then I'd test those searches across AI systems and inspect the sources being surfaced. If query fan-out is visible, I'd look at that too. Which sites appear repeatedly? Which competitors are consistently recommended? What language is being used to describe them?

That tells you where to focus.

You may need better solution pages. You may need stronger industry pages. You may need comparison content. You may need to improve your presence on a third-party platform. You may need more legitimate reviews. Or you may need to address a reputation problem that's being reflected back at you by the AI.

Finally, I'd make sure the measurement goes beyond mentions and citations. Track them, but connect them to leads, opportunities and revenue wherever you realistically can. Self-reported attribution is a simple place to start.

Good SEO is still good GEO

AI is changing search. I'm not arguing otherwise.

Query fan-out matters. Personalization matters. AI systems can synthesize information from multiple sources before producing an answer, and that changes how marketers need to think about brand visibility.

But we don't need to pretend that every tactic is new.

If somebody tells you to build useful pages, improve internal linking, create demand, strengthen your reputation, earn third-party mentions and make sure the web clearly understands what your company does, they're describing many of the same things good SEOs have cared about for years.

The technology is changing. The interface is changing. The customer journey is changing. Some of the tactics will change too.

The fundamentals haven't disappeared.

So before you rebuild your entire marketing strategy around GEO, AEO or whatever acronym comes next, get the basics right.

Then study what the AI systems are actually doing.

That's where the genuinely new opportunities start.

Work with Harpreet
Want to improve your presence in AI Search?
Book a free consultation and let's talk through your organic search strategy.
Book a free consultation →