Key takeaways
- Core technical SEO skills (site structure, crawlability, page speed) remain essential because AI models still rely on well-structured, indexable content to form their answers.
- Keyword stuffing, exact-match optimization, and chasing position #1 rankings are largely obsolete tactics in an era of AI Overviews and zero-click search.
- The skills gaining value are topical authority, entity-based content, structured data, and understanding how AI models cite sources.
- Google still holds 79.1% of the global search market, so traditional SEO isn't dead -- it's just no longer sufficient on its own.
- Generative Engine Optimization (GEO) is the new layer on top of SEO: getting your content cited by ChatGPT, Perplexity, Claude, and Google AI Overviews requires different thinking than ranking in a blue-link SERP.
There's a version of this conversation that's been happening in every marketing Slack channel for the past two years: "Is SEO dead?" The honest answer is no -- but the more useful answer is that the question itself is slightly wrong.
The real question is: which SEO skills still move the needle, and which ones are you wasting time on?
AI Overviews now appear on a huge share of Google searches. ChatGPT has become a research tool for millions of users who never open a search results page. Perplexity, Claude, and Gemini are answering questions that used to send people to your blog. The traffic patterns have changed. But the underlying logic -- that being findable, credible, and useful online matters -- hasn't gone anywhere.
What has changed is the specific skills required to execute on that logic.
Skills that still matter (and why)
Technical SEO fundamentals
This one surprises people, but technical SEO is arguably more important now than it was five years ago. Here's why: AI models don't just pull from a magical knowledge base. They crawl the web, index pages, and retrieve content. ChatGPT uses Bing. Perplexity runs its own crawler. Google's AI Overviews pull from the same index that powers traditional search.
If your site has crawl errors, broken internal links, slow load times, or thin pages that get de-indexed, AI models simply won't find your content. The foundation hasn't changed -- a site that search engines can't read is a site that AI can't cite.
Skills that remain directly relevant:
- Crawl budget management and fixing crawl errors
- Core Web Vitals and page speed optimization
- XML sitemaps and robots.txt configuration
- Canonical tags and duplicate content resolution
- Structured data / schema markup (more on this below)
Tools like Screaming Frog are still the go-to for crawl audits.
Structured data and schema markup
If there's one technical skill that's gotten more valuable in the AI era, it's schema markup. Structured data helps AI systems understand what a page is about without having to infer it from prose. FAQ schema, HowTo schema, Article schema, Product schema -- these are signals that AI models use to extract and cite information accurately.
When you mark up your content properly, you're essentially making it easier for AI to quote you. That's a direct competitive advantage.
Topical authority and content depth
Old SEO rewarded pages that targeted individual keywords. Modern SEO -- and AI search optimization -- rewards sites that own a topic. If you publish 40 shallow posts that each target one keyword, you'll get outcompeted by a site that publishes 15 deeply researched pieces that cover a topic from every angle.
AI models are trained to recognize expertise. When Perplexity or ChatGPT answers a question about, say, B2B SaaS pricing models, it doesn't just grab the page that ranks #1 for "B2B SaaS pricing." It synthesizes from sources it considers authoritative on the broader topic. Building that kind of depth is a content strategy skill, not a keyword skill.
Understanding search intent (at a deeper level)
Intent has always mattered in SEO. But the bar has risen. AI search handles nuanced, conversational queries that traditional search engines struggled with. Someone asking "what's the difference between a sole trader and a limited company if I'm a freelance designer in the UK" is a real query that AI handles well -- and if your content genuinely answers that specific question, you can get cited.
The skill here is moving from "what keyword is this person searching?" to "what problem is this person trying to solve, and what do they need to know to solve it?" That's a harder question, and it produces better content.
Link building (with caveats)
Backlinks aren't dead, but the way they work has shifted. Links from authoritative, topically relevant sites still signal credibility to both Google and the AI systems that use Google's index. What's changed is that mass link acquisition schemes, guest post farms, and low-quality directory links have become even less effective -- and potentially harmful.
The skill that matters now is earning genuine editorial links by publishing content worth citing. That's harder to scale, but it's also more durable.
Brand building and entity optimization
AI models have a concept of entities -- people, companies, products, concepts -- and they build associations between them. If your brand is consistently mentioned alongside certain topics across the web (including Reddit, YouTube, and news sites), AI models develop a stronger association between your brand and those topics.
This means brand building is now an SEO activity. Getting mentioned in industry publications, building a Wikipedia presence, being cited in discussions on Reddit -- these all feed into how AI models perceive your brand's authority.
Skills that AI search has made largely irrelevant
Exact-match keyword density optimization
The practice of hitting a specific keyword density -- "use your target keyword 12 times per 1,000 words" -- was already losing relevance before AI search. Now it's genuinely counterproductive. AI models evaluate semantic meaning, not keyword frequency. Content written to hit a keyword count reads poorly and gets deprioritized by both Google's helpful content systems and AI citation algorithms.
Write for humans. Use natural language. Cover the topic thoroughly. That's the entire strategy.
Chasing position #1 for informational queries
For informational queries -- "how does X work," "what is Y," "best way to Z" -- the traditional goal of ranking #1 in the blue-link results is increasingly beside the point. Google's AI Overviews now appear above organic results for a huge share of these queries. The user gets their answer without clicking.
This doesn't mean you should abandon informational content. It means the goal has changed: you want to be the source that AI Overviews cite, not just rank below them. That requires different optimization.
Meta keyword tags
This one has been dead for years, but it's worth saying explicitly: meta keywords do nothing. If you're still filling them in as part of a workflow, stop.
Exact-match domain strategies
Buying a domain like "best-plumbers-london.co.uk" to rank for "best plumbers London" used to work. It doesn't anymore, and it signals low quality to both Google and AI systems. Domain authority now comes from brand recognition and genuine editorial links, not keyword-stuffed URLs.
Thin content at scale
The old playbook of publishing hundreds of short, keyword-targeted pages -- each one covering a slightly different variation of the same query -- is actively harmful now. Google's helpful content updates have penalized this approach, and AI models won't cite thin pages. If you're still running a content operation built around volume over depth, this is the year to change that.
Manual rank tracking as a primary KPI
Tracking your position for 50 keywords in Google's blue-link results used to be the core metric of SEO success. It's still useful context, but it's no longer the primary signal. A brand can rank #3 for a query but never appear in the AI Overview that sits above all organic results. Conversely, a page that ranks #8 might get cited constantly by ChatGPT.
The metric that matters now is AI visibility -- how often your brand and content appear in AI-generated answers across the platforms your customers actually use.
The new skills you need to develop
Generative Engine Optimization (GEO)
GEO is the practice of optimizing content so that AI models cite it. It overlaps with traditional SEO but requires additional thinking: How does this content answer a specific question? Is it structured in a way that's easy for AI to extract? Does it include the kind of factual, specific information that AI models prefer to cite?
Concrete GEO tactics include writing direct, declarative answers to questions (not burying the answer in paragraph three), using clear headings that match how people phrase queries, and including specific data points and examples that AI models can quote.
Understanding AI crawler behavior
AI companies run their own web crawlers. Knowing which crawlers are hitting your site, which pages they're reading, and whether they're encountering errors is genuinely useful intelligence. If the Perplexity crawler keeps hitting a 404 on your most important product page, that's a problem worth fixing.
Promptwatch has a crawler logs feature that shows exactly which AI crawlers are visiting your site, which pages they read, and how often they return -- something most traditional SEO tools don't cover at all.
Prompt and query research
Traditional keyword research focused on search volume and competition in Google. The equivalent for AI search is understanding which prompts people use when talking to ChatGPT, Perplexity, or Claude -- and which of those prompts your competitors are appearing in but you're not.
This is a genuinely new research skill. The queries are longer, more conversational, and often more specific than traditional search queries. Tools built for AI visibility tracking can surface these prompt patterns and help you prioritize which content gaps to close first.
Content gap analysis for AI citations
The most actionable new skill in 2026 is identifying the specific questions that AI models answer using your competitors' content instead of yours -- and then creating content that fills those gaps. This is different from traditional keyword gap analysis because you're not just looking at what ranks; you're looking at what gets cited in AI-generated answers.
Multi-platform visibility thinking
Your customers aren't just using Google. They're asking ChatGPT, using Perplexity for research, getting recommendations from Gemini, and seeing AI Overviews in their search results. An SEO strategy that only optimizes for one platform is leaving visibility on the table.
The skill here is understanding how different AI models behave, what they tend to cite, and how to optimize for each without creating completely separate content strategies (because the fundamentals overlap more than they differ).
A comparison: old SEO vs. AI-era SEO
| Skill / tactic | Still relevant? | Why |
|---|---|---|
| Technical site health (crawlability, speed) | Yes, more than ever | AI crawlers need accessible, fast pages |
| Schema markup / structured data | Yes, increased value | Helps AI extract and cite content accurately |
| Topical authority building | Yes, core strategy | AI models prefer recognized subject-matter experts |
| Link building (quality) | Yes, with caveats | Still signals credibility to Google and AI |
| Keyword density optimization | No | AI evaluates semantic meaning, not keyword count |
| Exact-match domains | No | No ranking benefit; signals low quality |
| Chasing #1 for informational queries | Mostly no | AI Overviews sit above organic results |
| Thin content at scale | No | Penalized by Google; ignored by AI |
| Meta keywords | No | Has been irrelevant for years |
| Manual rank tracking as primary KPI | Partial | Useful context but not the main signal anymore |
| AI citation / GEO optimization | New skill | The new layer on top of traditional SEO |
| Prompt and query research | New skill | Understanding how people query AI models |
| AI crawler log analysis | New skill | Know which AI bots visit your site and what they read |
How to audit your own skill set
The practical question is: where are you spending your time, and is that time well-invested given how search has changed?
A few diagnostic questions worth asking:
- Are you still optimizing for keyword density rather than topical depth? If yes, that's time better spent elsewhere.
- Do you know which AI models are citing your content, and for which queries? If not, you're flying blind on a significant and growing traffic channel.
- Is your content structured to answer specific questions directly, or is it written to rank for keywords? The former gets cited; the latter increasingly doesn't.
- Are you tracking AI crawler activity on your site? If AI bots are hitting errors on your key pages, that's a fixable problem -- but only if you know it's happening.
For teams that want to get a clearer picture of their AI visibility, platforms like SE Ranking cover traditional SEO alongside AI visibility tracking.
For deeper AI-specific analysis, tools like Semrush have added AI Overview tracking to their existing SEO suites.
And for content optimization -- making sure your writing is structured in a way that both humans and AI models find useful -- Clearscope and Surfer SEO remain solid options.
The bottom line
SEO in 2026 isn't a different discipline -- it's an expanded one. The technical fundamentals still matter. Content quality still matters. Links still matter. But the finish line has moved: it's no longer just about ranking in a list of blue links. It's about being the source that AI models trust enough to cite when your customers are asking questions.
The skills that are becoming obsolete are mostly the ones that were always shortcuts -- tactics that worked because of how algorithms were built, not because they genuinely served users. Those shortcuts are closing fast.
The skills gaining value are the ones that were always the right answer: deep expertise, clear writing, genuine authority, and a real understanding of what your audience needs to know. AI search has just made it harder to fake those things.