AI Is Evolving SEO, Not Killing It
I have been doing SEO since 2013.
That gives me more than a decade of watching “this will kill SEO” headlines march across the internet like tiny digital zombies.
Mobile search was going to kill SEO.
Voice search was going to kill SEO.
Featured snippets were going to kill SEO.
ChatGPT launched in November 2022, and the panic cycle started again.
SEO is still here.
The search experience keeps changing. The work keeps getting more interesting. Businesses still need to be found when people have questions, compare options, choose a provider, visit a store, or pull out a credit card.
AI is changing how people discover information. That changes the job. It also gives good SEO work more places to matter.
The sky is falling again
Every major search change creates the same reaction.
People see a new feature, imagine the worst possible outcome, and start writing the obituary for an entire industry.
The current version says AI will answer every question, eliminate website clicks, and make search engine optimization irrelevant.
That story has one major weakness. People still need information from businesses.
A dispensary still needs to show up when someone searches for a nearby store. A packaging supplier still needs to explain its products to a buyer with a long sales cycle. A roofing company still needs service pages, location signals, reviews, and a clear way to request an estimate.
The format of discovery can change while the business need stays familiar.
I have watched clients move through this process many times. They ask Google a question, read an educational page, compare providers, check reviews, visit a location, and contact the business.
AI can influence several steps in that journey.
It does not remove the journey.
What AI actually changed
AI changed the search surface first.
Google AI Overviews can place a generated summary above traditional organic results. AI Mode gives people a more conversational way to explore a topic, ask follow-up questions, and refine a search without starting from scratch every time.
ChatGPT, Perplexity, Gemini, Claude, and Copilot add more places where people can ask for recommendations, comparisons, explanations, and sources.
That means a company can be discovered through:
- A standard organic result
- An AI Overview citation
- A follow-up in AI Mode
- A recommendation in a conversational answer
- A local business profile
- A product result
- A review or discussion on another platform
The page underneath still matters.
AI systems need information to summarize. Search engines need pages to crawl and index. People need somewhere to verify details, explore products, learn about a service, or take the next step.
The work now falls under what I call Search Everywhere Optimization.
That means traditional SEO, answer engine optimization, and generative engine optimization working as one program. I have written more about AEO and GEO and the value of those labels because the terminology matters less than the actual work.
The evidence is more complicated than the panic
AI answers can reduce clicks on some informational searches.
That deserves attention. It also deserves context.
An Ahrefs study published on May 19, 2025 analyzed 55.8 million AI Overviews across 590 million searches. The study reported AI Overviews appearing across 9.46% of desktop keywords in its index and 16% of US desktop keywords. It also estimated that at least 12.8% of searches by volume showed an AI Overview in the data reviewed.
Those numbers describe a large and changing search feature. They do not describe every query, industry, device, or customer journey.
The same Ahrefs study found that AI Overviews appeared more often for informational, longer, non-branded searches. Those searches often help people learn before they buy. They also tend to have lower direct conversion value than a search for a specific service, product, location, or brand.
Semrush found a similarly messy picture in its refreshed AI Overviews study, published December 15, 2025. AI Overviews appeared for 6.49% of tracked queries in January 2025, peaked at 24.61% in July, and stood at 15.69% in November.
When Semrush and Datos tracked the same keywords before and after an AI Overview appeared, the zero-click rate moved from 33.75% to 31.53%.
That finding does not mean AI Overviews help every website. It shows that user behavior depends on the query, the intent, the quality of the answer, and the options presented around it.
Search behavior is more complicated than one universal click-loss number.
What changed since the original version
When I first wrote this post in January 2026, AI search was already moving quickly.
Since then, AI features have become more familiar inside everyday search. People do not need to visit a separate AI tool to encounter an AI-generated answer. Google can place an overview directly in the results, and AI Mode gives the interaction more room to expand.
That changes the first impression a searcher gets.
A business can rank well and still have its result pushed lower on the page by an answer layer. A business can also earn visibility through a citation in that answer. The two experiences can exist together.
AI Mode adds another wrinkle because the user can keep asking questions. A first search may ask for a definition. The next may ask for examples. The next may ask for local options or a comparison.
That makes clear entities and well-connected content more important.
Google’s official guidance on AI features in Search points site owners toward familiar foundations. Pages need to be crawlable, indexable, useful, internally linked, and supported by accurate structured data that matches visible content.
The search interface got more conversational.
The website still needs to make sense.
The SEO winners are adapting
The businesses doing well in this environment are using AI to strengthen human expertise.
They use AI for research, question discovery, outline development, content gap analysis, and workflow support. A person still checks the facts, makes strategic decisions, reviews claims, and decides what deserves publication.
I use AI tools in my own work. They help me move faster through research and spot patterns earlier.
I also reject plenty of AI-generated suggestions because they are vague, repetitive, inaccurate, or completely disconnected from the client’s actual business.
Unfiltered output does not become useful because it came from an expensive model.
The strongest sites are building depth around subjects they genuinely understand. My guide to building topical authority explains how connected pages give search systems more evidence about a business, its expertise, and the topics it covers.
That work includes:
- Clear service, product, and location pages
- Educational content that answers real questions
- Internal links that explain relationships between pages
- Consistent business and product information
- Accurate author and organization details
- Sources that support important claims
- Content written for the customer’s actual decision process
A page should answer the question before it asks for the sale.
That principle has worked for me in cannabis SEO, local SEO, B2B, e-commerce, health and wellness, and other industries where buyers research heavily before making a decision.
How I approach SEO now
I start with the business and the customer.
What does the company sell? Who needs it? What questions come up before a purchase? What information is difficult to explain? What claims need extra review? What action matters to the business?
Then I look at the site.
I review technical issues, search demand, existing rankings, internal linking, indexed pages, content gaps, conversion paths, and local signals. I want to know what the site already communicates and where the message becomes muddy.
From there, I build a content plan that connects the research journey to the business.
My post about building a content SEO strategy that AI search uses covers that process in more detail. The short version involves matching each question to the right page type.
A definition may need an educational guide. A city-based query may need a local landing page. A product comparison may need a buying guide. A high-intent search may need a service or product page with process details, proof, and a clear next step.
I also make the answer easy to find.
Important information should not hide beneath three paragraphs of generic introduction. Headings should describe the question. Sections should stay focused. Definitions should be clear enough to stand on their own.
That helps readers skim.
It also gives search systems cleaner passages to interpret.
The real threat is refusing to adapt
McKinsey’s State of AI 2025 report, published November 5, 2025, found that 88% of respondents say their organizations use AI in at least one business function.
The same report found that most organizations are still in experimentation or piloting stages with AI.
That number tells me something important about the market.
Clients, competitors, customers, and coworkers are already experimenting with AI. The businesses that ignore the shift may lose opportunities to competitors who are learning how search behavior is changing.
Refusing to adapt creates a bigger problem than AI itself.
Adaptation does not require throwing out every existing SEO process. It requires reviewing the process with fresh eyes.
I still care about crawlability, page structure, internal links, search intent, local relevance, technical health, and conversion paths. I also care about how clearly a business is represented across the sources AI systems may use.
That includes the website, business listings, reviews, directories, social profiles, industry mentions, product data, and other trusted references.
The information needs to agree.
A search system has an easier job understanding a company when the company describes itself consistently across the web.
Set realistic expectations
AI answers will reduce clicks for some informational queries.
The impact varies by industry and query type. A person researching a broad definition may get enough information from an overview. A person comparing dispensaries, contractors, products, or suppliers still needs details that support a real decision.
The best opportunities usually sit where expertise, intent, and business value overlap.
Businesses that perform well in this environment tend to have clear entities, real expertise, useful supporting content, accurate information, and pages that answer a question before selling something.
I cannot promise rankings, citations, traffic, or revenue on a fixed schedule. Search systems make their own decisions, and AI outputs change based on the platform, query, location, freshness, and available sources.
I can build a stronger foundation and measure what happens.
That includes rankings, organic traffic, branded searches, calls, form fills, store visits, qualified leads, revenue-related actions, and AI visibility when reliable tracking is available.
Before investing in a shiny new acronym, start with a free SEO audit and find out what the existing site needs.
Where SEO is heading
SEO is moving into more search environments.
Google still matters. AI answer engines matter. Local platforms, shopping systems, video platforms, review sites, and industry directories can all influence discovery.
The brands with the strongest chance of being found will make their information easy to crawl, understand, retrieve, verify, and act on.
That takes strategy and implementation.
It takes content that sounds like a person who knows the subject. It takes technical work that keeps the site accessible. It takes internal links that connect the answers. It takes human review, especially in regulated and claim-sensitive industries.
AI is changing the rules around attention.
It is also making clear why useful content, strong entities, and real expertise matter.
SEO has survived every previous search disruption because people still need answers and businesses still need customers.
The work is in motion.
Feeding the algorithm since 1997.
Make Randi One of your Preferred Sources