Is SEO Dead in the AI Era? How to Run SEO and GEO Together

Every few months a client forwards me an article with the same headline: SEO is dead. This year the argument feels stronger, because AI answers now sit on top of the results page. My short answer is still no. However, the old version of SEO really has run its course. In this guide I explain how I run SEO and GEO as one plan instead of two competing projects, and I back the argument with published data rather than opinion.

Is SEO dead, or should you run SEO and GEO together?

SEO is not dead; it has changed shape. Search engine optimization is still the work of making pages crawlable, trustworthy and genuinely useful. When you run SEO and GEO together, the same content can rank in Google and also appear as a cited source in ChatGPT, Gemini or AI Overviews. In other words, the foundation stays shared.

In practice, most of the work does not change. Technical health, content quality and brand trust decide visibility in both channels. What does change is how you format answers, how you measure progress and how you define success. So the useful question is not "SEO or GEO?" but "how do I fold both into one roadmap?" If you need the definition first, read my explainer on generative engine optimization (GEO). Here I skip the theory and focus on the operating plan.

Why does the "SEO is dead" debate come back every year?

I have heard this story before. People said it when social media took off, when mobile search exploded and when voice assistants arrived. Each time search behavior shifted, but the need to find information never went away. That said, the AI wave is bigger than the earlier ones. This time the answer often appears on the results page itself.

There is also a commercial reason the debate never dies. For example, a dramatic headline gets clicks, sells new tools and packages new services. On top of that, some sites really are losing traffic, and that loss fuels the fear. I look at the question through two simple tests:

• Do users still want to reach a source, a brand or a product?
• Which signals does an AI answer rely on when it chooses that source?

So both answers lead back to core SEO. AI search tools crawl, read and select web pages before they write a response. In other words, visibility still depends on content that machines can access, understand and trust.

What does the data actually say about clicks and AI answers?

I prefer to tie both the panic and the reassurance to data. A Pew Research Center study published in July 2025 found that Google users who saw an AI summary clicked a traditional result in 8% of visits. By contrast, users who did not see a summary clicked in 15% of visits. Clicks on links inside the summary happened in just 1% of visits.

Google, on the other hand, tells a different story. The company reported that AI Overviews reached 2 billion monthly users in the second quarter of 2025. In addition, Head of Search Liz Reid wrote in an August 2025 blog post that total organic click volume had stayed "relatively stable" year over year. Still, the post did not include a detailed dataset to support the claim.

On the ChatGPT side, OpenAI announced 800 million weekly active users in October 2025. In short, people are not searching less; their searching is simply spread across more surfaces. As a result, informational clicks shrink, while a brand mention inside an answer becomes a new form of visibility.

What does Google say about SEO for AI features?

Google's official position is refreshingly direct. Its Search Central page on AI features and your website states that SEO best practices remain relevant for AI Overviews and AI Mode. According to the same page, you need no additional requirements or special optimizations to appear there.

The page also explains a useful detail. AI Overviews and AI Mode may use a "query fan-out" technique, which means they run several related searches across subtopics before they build a response. Therefore pages that answer the follow-up questions around a topic have an edge over pages that target one keyword. Google also notes that traffic from AI features counts toward overall search traffic in Search Console.

Meanwhile, the familiar controls still apply. If you want to limit how your content appears, you can use nosnippet, data-nosnippet, max-snippet or noindex. Google-Extended manages whether Google can use your content to train Gemini models, and Google says it does not affect ranking in Search. So from Google's point of view, SEO and GEO share one foundation with two presentation layers.

What changes and what stays the same?

I usually show clients this split in a table. Panic tends to come from the belief that everything has changed. In reality, the changing part is specific and manageable.

The takeaway is simple: the core stays, the edges move. Consequently, adding a GEO layer to your existing SEO process is far more efficient than splitting budget and people into two teams.

Here is how that looks on a single page. When I rewrite a service page, I check three things in one pass. First, does it rank for the right query? Next, does the opening paragraph answer the question on its own? Finally, does the brand name appear naturally inside that answer? A separate GEO project would send the same page through two teams at two different times. As a result, the cost doubles and the tone of the page usually suffers.

Is GEO replacing SEO or building on top of it?

I see GEO as a layer, not a new building. AI search systems usually rely on a search index when they assemble an answer, and Google's own AI surfaces draw on the Google index. Therefore a page that search engines cannot crawl, index or trust has little chance of becoming a source in an AI answer.

Being a layer does not make it trivial, though. A brand stuck on page two in classic results may never appear in an AI answer at all. Moreover, even a page that ranks well can lose the citation to a competitor if it buries the answer. For a framework on direct answer writing, see my guide to answer engine optimization (AEO).

Put simply, GEO extends the job SEO already does into AI surfaces. Adding GEO tactics to a site with a weak foundation is like painting a cracked wall.

How do you merge SEO and GEO into one operating plan?

In my consulting work I use one roadmap. I tag every task by its impact on both channels. That way the team never feels it is running two projects, and nobody does the same work twice. The plan has five steps:

1. Foundation audit: check crawlability, indexing, speed and structured data.
2. Question map: collect the real questions buyers ask before they purchase.
3. Content restructuring: give every critical page a clear answer block.
4. Brand signals: publish consistent, verifiable brand facts across the web.
5. Unified measurement: track rankings, clicks, AI mentions and conversions in one report.

The order matters, because each step depends on the one before it. For example, if you rewrite content before you set up measurement, you will never know what worked. Likewise, if the technical base is broken, even the best answer block goes unread. That is why I start every project with the audit, and why the first month of my SEO consulting engagements usually goes into it.

Which fundamentals serve both channels at once?

If your budget is tight, invest first in work that pays off twice. From what I see in client accounts, these tasks support Google rankings and AI visibility at the same time:

• Clean, server rendered HTML that does not depend on JavaScript.
• A heading structure where each page answers one main question clearly.
• Accurate structured data such as Organization, Product, Service and FAQPage.
• Content with a named author, a visible date and cited sources, kept up to date.
• Topic clusters with internal links that connect them.
• Fast pages that work smoothly on mobile.

Admittedly, none of this is new. However, AI surfaces punish gaps in these basics harder than before. If a model struggles to understand your page, it simply picks a competitor that states the same fact more clearly. You can draft markup quickly with my schema generator.

What should your content team write differently?

The biggest shift I see is on the content side. A few years ago, a long guide with the keyword in the right places was often enough. Now that guide also needs quotable paragraphs that make sense on their own.

My practical rule is this. Under each main heading, place a 40 to 60 word paragraph that answers the question directly. Then move on to detail, examples and exceptions. As a result, readers find the answer fast, and an AI system can lift that paragraph without losing its meaning.

Original data and experience also matter more than ever. After all, a language model already knows the generic facts. To earn the citation, you need to offer something it cannot find elsewhere: your own field observations, your own calculation method, your own comparison. I also collected the writing rules in my guide on how to write content for AI Overviews.

Finally, keep your terminology consistent. If you call the same service by three different names across your site, a model will have a hard time treating your brand as one clear entity.

Which technical checks matter most now?

The basics of technical SEO have not changed, but my checklist has grown. The first place I look is robots.txt. Many sites block every AI bot in one sweep and accidentally shut out the crawlers that power AI search as well.

OpenAI, for instance, separates two bots in its crawler documentation. OAI-SearchBot crawls pages so they can appear in ChatGPT search results, while GPTBot collects data for model training. You can block one and allow the other. So a company that wants to stay out of training data but still appear in ChatGPT search should block only GPTBot.

These are the items I have added to my technical checklist:

• Decide separately for search crawlers and training crawlers in robots.txt.
• Confirm that key content appears in the HTML without JavaScript.
• Watch server logs to see which pages AI crawlers actually visit.
• Check that firewall or CDN rules do not silently block these bots.

My robots.txt generator helps you write the rules correctly. For a deeper look, read my article on technical SEO after AI.

Why do brand signals matter more than before?

In classic SEO, links were the main signal of authority, and they still count. In AI answers, however, the way your brand appears across the web also carries weight. A model tends to recommend brands that show up consistently in many independent sources.

That is why I ask clients to keep brand facts identical everywhere. Your company name, service description, address, founding year and area of expertise should match on your site, your Google Business Profile, industry directories and press coverage. In addition, mentions on independent sites, forums and comparison articles help a model link your name to a category.

To understand which brands tend to win, read my analysis of which brands AI search engines recommend. Then test your own position with the method in how your brand shows up in ChatGPT and Gemini.

How do you measure SEO and GEO performance?

Measurement is the weakest link right now. Google includes AI Overviews and AI Mode traffic in the general web search data in Search Console, but it offers no separate filter. In other words, you cannot cleanly split AI surface clicks from classic result clicks today.

So I build measurement in three layers. First, Search Console: I track changes in impressions, clicks and click through rate by query type. Second, GA4: I group referral traffic from chatgpt.com, perplexity.ai and gemini.google.com into a dedicated channel. Third, a manual mention test.

For that test, I list the 20 to 30 questions that matter most in the client's buying journey. Every month I ask them in several AI tools. Then I log whether the brand appears and which sources the tools cite. The method is not perfect, because answers vary by user and by day. Even so, it gives a reliable enough compass for the trend.

To keep campaign links separate in your reports, use my UTM builder.

Which KPIs no longer work on their own, and which should you add?

Rank tracking alone can now mislead you. A page can hold position one while an AI summary above it absorbs the clicks. Consequently, you will often hear the question: "Our rankings are the same, so why did traffic drop?"

In my reports I now track these indicators side by side:

• Impressions and click through rate, split by informational and commercial queries.
• Changes in branded search volume.
• Referral traffic from AI tools and its conversion rate.
• The share of test questions where the brand appears.
• Leads and sales from the organic channel.

Of these, branded search is the one I watch most closely. A user who sees your name in an AI answer often does not click, but may search for your brand a few days later. Classic attribution models miss that effect. That is also why I never judge a channel without looking at the bottom line: leads and revenue.

How should you split budget between SEO and GEO?

There is no universal formula, and you should be wary of anyone who offers one. Still, I use a starting point in client conversations. The ranges below are a starting range based on my field experience, not a guarantee.

On a site with a weak technical base, I put most of the budget, roughly two thirds, into classic SEO foundations, because that work benefits both channels. On a site with solid foundations, you can shift more toward GEO work such as answer rewriting, brand mentions and measurement. In my experience that share rarely exceeds half of the total.

Example calculation: imagine a company with a monthly consulting and content budget of 4,000 US dollars and a healthy technical setup. It might put 2,500 dollars into technical upkeep and new content, and 1,500 dollars into answer focused rewrites and mention testing. These figures only illustrate the ratio; your numbers will differ.

I discuss the broader budget logic in organic growth or paid ads.

How should you audit the content you already have?

Before you create anything new, look at what you already publish. On most sites, the real opportunity sits in pages that already rank but deliver the answer too late. I start the audit with Search Console data and sort pages into three groups:

1. High impressions with a falling click through rate: update the answer block and the title.
2. Close to a sale but ranking poorly: add internal links, structured data and depth.
3. No traffic and outdated: merge them or remove them.

Then run each important page through a simple test. Read the first paragraph on its own and ask yourself whether it answers the question without any context. If it does not, an AI system will struggle to quote it, and so will a human reader.

This audit usually takes a few days, yet it sets the priorities for the months ahead. My readability checker helps you spot paragraphs that run too long.

What should you do in the first 90 days?

When I start a new project, I break the first quarter into concrete milestones. That way the team knows what to do, and leadership can see progress. Here is the schedule I recommend:

1. Month one: technical audit, robots.txt and crawler access review, Search Console and GA4 validation, first mention test.
2. Month two: answer focused rewrites of the ten highest revenue pages, structured data, a complete question map.
3. Month three: aligned brand facts on external sources, new question content, a second mention test and a comparison report.

Do not expect miracles after 90 days. With solid measurement, though, you can see which pages start to appear in AI answers and which queries show a change in click through rate. That evidence is enough to set the next quarter's priorities.

Be especially patient in month two. Google and AI tools both need time to reassess rewritten pages.

Which mistakes waste SEO and GEO work?

The mistakes I see in the field are surprisingly similar. Most come from chasing the new channel and neglecting the foundation. These are the most expensive ones:

• Blocking every AI crawler in robots.txt without thinking it through.
• Mass producing AI content with no original value.
• Reading ranking reports while ignoring clicks and conversions.
• Pinning your hopes on a single file such as llms.txt.
• Leaving contradictory brand facts across platforms.
• Rewriting pages that already rank well and breaking them.

The second point deserves emphasis. Google's spam policies explicitly target low value content that anyone produces at scale, whether with AI or by hand. So the idea that "content is cheap now, let's publish a thousand pages" hurts both SEO and GEO. Fewer pages that truly help people perform better in both channels.

How do zero click searches affect your strategy?

The most visible effect of AI summaries is the rise of zero click searches. A user gets the answer on the results page and leaves. You can see this most clearly in definitions, conversions and simple fact queries.

Not every query behaves the same way, however. For searches close to a purchase, comparisons or local services, people still visit websites. That is why I recommend splitting your content strategy by query type. For informational queries, aim for brand visibility. For commercial queries, aim for visits and conversions.

Once you make that split, the traffic loss looks less scary. Many of the lost clicks were visits that would never have turned into sales anyway. I cover query types in more depth in what zero click searches are.

Where does Google Ads fit into this picture?

When organic visibility becomes less predictable, paid search acts as a controllable buffer. I plan SEO and GEO work alongside Google Ads, because the two channels feed each other with data.

For example, the search terms report in your campaigns shows the exact words customers use, which enriches your question map. In the other direction, topics that perform well organically are safer bets for paid promotion. Google also shows ads within and around AI Overviews, so the paid channel has a place on these new surfaces too.

Just do not treat ads as a replacement for organic work. When the ad budget stops, the traffic stops with it. Only a solid organic foundation builds lasting visibility. In my Google Ads management service, I build the paid structure together with the organic strategy.

Which businesses should act first?

Not every business faces the same urgency. Publishers, blogs and guide sites that rely mainly on informational content feel the impact fastest. These businesses need to rethink their model.

B2B service firms, SaaS companies and consultancies come next. Their buyers increasingly turn to AI tools when they build a shortlist. As a result, missing from answers to questions like "best providers of X" quietly costs them deals.

Local service businesses and ecommerce stores also feel the shift, but transactional queries still produce strong clicks. For this group, the priority should be structured, consistent product and service information. Whatever group you belong to, measure your current visibility before you build a strategy.

The verdict: SEO is not dead, it has evolved

SEO is not dead in the AI era, but the "rank and collect clicks" model no longer works on its own. Your goal now is to become a trusted source for the questions your customers ask, both in search results and in AI answers. The way to get there is not two separate teams. Instead, it is one approach that merges SEO and GEO into a single plan.

To sum up: keep the foundation solid, write quotable content, keep your brand facts consistent and update your measurement for the new reality. If you want a second opinion on where your site stands today, reach out through my contact page, and we can agree on the first step together.