There are now two audiences reading every page you publish: humans and machines.
Google’s crawlers still evaluate your title tags, keyword placement and link structure the same way they have for years. But alongside them, large language models (LLMs) like ChatGPT, Perplexity, Claude and Google’s own AI Overviews (and Gemini) are now scanning your content to decide whether it’s worth citing in the answers they generate for millions of users.
The good news is that most on-page optimizations that work for traditional search also work for AI search. Clear structure, strong headings, direct answers and authoritative content perform well in both contexts. The differences come down to emphasis – LLMs reward certain formatting habits and trust signals more heavily than Google’s traditional algorithm does.
This guide covers everything you need to know about on-page SEO in 2026, from the fundamentals that still matter to the newer practices that make your content visible to AI platforms. Whether you’re optimizing a service page, a blog post or an entire site, these are the things that actually move the needle and the things that our SEO experts in Canada implement on a day-to-day basis.
How AI Search Changes On-Page SEO
Before getting into tactics, it’s worth understanding how AI search platforms consume your content – because it’s fundamentally different from how Google works.
When someone asks ChatGPT or Perplexity a question, the AI doesn’t usually paste the full query into a search engine. It typically breaks the question down into smaller sub-queries and searches for each one separately. So if a user asks “What’s the best CRM for a 10-person sales team on a tight budget?” the AI might run three separate searches: “best CRM small business 2026,” “CRM features sales teams” and “CRM pricing comparison.” It then reads the top results from each sub-query, combines the information and generates a single answer.
This has a direct implication for your content. You often don’t need to match the exact long-tail question a user types into ChatGPT. You need to rank well for the shorter sub-queries that the AI extracts from that question.
The other major difference: LLMs prioritize content they can extract and cite cleanly. A Search Engine Land study of 1.2 million verified ChatGPT citations found that 44% of all citations come from the first 30% of a page’s content. LLMs scan for the answer before deciding whether to cite the source. If your content buries the answer under too much preamble, you lose.
Traditional SEO is the foundation of AI search visibility. If your content isn’t indexed, AI tools can’t find it. Everything below strengthens your position in both systems simultaneously:
1. Title Tags
The title tag (the page’s headline in Google search results) is still the single most influential on-page element for traditional search rankings. Google’s own documentation confirms that the presence of relevant keywords in the title tag remains a “basic signal” for determining content relevance.
Keep your title tags under 60 characters so they don’t get truncated in search results. Place your primary keyword as close to the beginning as possible. Make each title tag unique across your entire site.
For AI search, your title tag serves a slightly different purpose. LLMs use it to quickly understand what a page is about and whether it’s relevant to the sub-query they’re evaluating. A clear, descriptive title tag that explicitly states the topic improves your chances of being selected as a citation source.
Avoid vague or clever titles. “Everything You Need to Know About X” tells an LLM nothing. “How to [Do Specific Thing]: Step-by-Step Guide” tells it exactly what the page covers.
2. Meta Descriptions
Meta descriptions don’t directly influence rankings – Google has confirmed this. But they do influence click-through rates, which are a signal Google pays attention to. A well-written meta description can be the difference between someone clicking your result or scrolling past it.
Keep them under 155 characters for desktop and 120 characters for mobile. Include your primary keyword (Google will bold it when it matches the search query). Write them as a compelling pitch for the page, not a keyword-stuffed summary.
For AI search, meta descriptions are less important because LLMs ignore them and LLM users don’t see your meta descriptions either. But they’re still worth optimizing because higher click-through rates improve your traditional SEO rankings, which in turn increases the likelihood of your page being discovered and cited by AI platforms.
3. URL Structure
Clean, keyword-rich and well-structured URLs are a consistent ranking factor. They help users and search engines understand what a page is about at a glance.
Use hyphens to separate words (not underscores). Include your primary keyword. Keep URLs short and readable – strip out unnecessary words like “a,” “the” and “and.” Use lowercase only.
If you’re a plumber in Boston and want to rank for “drain cleaning boston”:
A good URL looks like: /drain-cleaning-boston
A mediocre URL looks like: /drain-cleaning
A bad URL looks like: /get-your-drains-cleaned-by-our-professional-and-friendly-plumbers
But the worst thing you can do is put all of your services on one page, without having a separate page dedicated to each one. It’s vital to have a separate page dedicated to each keyword that you want to rank for.
So if you’re a plumber in Boston, target “plumber boston” with your homepage and give a brief overview of each subservice on your homepage, but also have a separate in-depth page for each subservice.
4. Heading Structure (H1 to H6)
Your heading hierarchy is one of the most critical on-page elements for both traditional and AI search.
Use one H1 per page – it should closely match your title tag and clearly state the page’s primary topic. Use H2s for major sections and H3s for subsections within those. Don’t skip levels (going from H2 directly to H4 confuses both crawlers and LLMs). Include your primary keyword in the H1 and at least one H2 or H3.
Make your headings explicit and descriptive. Instead of “Our Approach,” write “How We Handle Kitchen Renovations in Toronto.” Instead of “Pricing,” write “Kitchen Renovation Costs in 2026.” The heading should make sense on its own, without requiring context from the rest of the page.
Question-based headings work particularly well for AI visibility because they mirror how users phrase queries. An H2 like “How Much Does a Kitchen Renovation Cost?” maps directly to the kinds of questions people ask LLMs.
5. Content That Works for Both Audiences
The biggest shift in on-page content strategy for 2026 is what you could call “answer-first” writing.
Traditional SEO content often follows a pattern of building up to the answer – providing context, background and supporting information before finally delivering the key takeaway (yes, those annoying recipe sites are a perfect example). That structure used to work well for traditional SEO but it doesn’t work well anymore, and it especially doesn’t work well for LLMs.
Lead every section with the direct answer or key point. Put the most important information in the first one to two sentences below each heading. Then provide the supporting detail, context and nuance afterward.
This doesn’t mean dumbing down your content. It means structuring it so that both a human reader scanning your page and an AI extracting your content can quickly find what they need. Think of it as writing in an inverted pyramid format – the same structure journalists have used for decades.
Beyond structure, here are the content principles that matter most in 2026:
Put your primary keyword in the first paragraph. Ideally the first sentence. This signals to both humans and algorithms that your content is relevant to the search query. It’s also important to mention the primary keyword and some similar variations a few other times throughout the body content.
Write in clear, direct language. LLMs process content more reliably when sentences are straightforward and fact-forward. Vague language gets passed over in favour of definitive statements.
Use specific data and cite your sources. When you make a claim, back it with a number, a study or a named source. The Princeton GEO study found that including citations and statistics can boost source visibility in AI-generated responses by over 40%. AI platforms want to cite content they can verify.
Cover topics comprehensively. A surface-level 500-word article on a complex topic will lose to a thorough 2,000-word guide every time – in both traditional and AI search. Build content clusters around core themes: create a comprehensive pillar page and then support it with related articles that link back to it.
Keep content current. Both Google and LLMs prioritize freshness. Update your high-performing pages regularly with new data, current examples and revised recommendations. Add visible “last updated” dates to signal freshness to both readers and machines and/or include the current year in the page title, headings and body content.
6. Keyword Strategy in 2026
Keywords still matter – but the way to use them has evolved.
Google’s algorithm has gotten sophisticated enough to understand synonyms, related terms and topical context. You still need your primary keyword in the title tag, H1, URL and naturally throughout the body content. But you also need to incorporate semantically related terms that signal topical depth.
If your primary keyword is “kitchen renovation costs,” your content should also naturally include terms like “budget,” “contractor pricing,” “material costs,” “labour,” “renovation timeline” and “project scope.” These related terms help search engines understand that your page covers the topic thoroughly, not just superficially.
For AI search, the concept is similar but with an added layer. LLMs build understanding through entity recognition – they look for named things (people, places, products, concepts) and the relationships between them. The more clearly you define and connect entities in your content, the easier it is for AI platforms to extract and cite your information.
Use your brand name explicitly rather than saying “our platform” or “the tool.” Mention specific products, methodologies and frameworks by name. When referencing other companies, tools or concepts, use their full proper names. This kind of entity-rich writing makes your content more parseable for AI systems.
7. Internal Linking
Internal links serve two purposes: they help search engines discover and understand the relationship between your pages, and they distribute link equity throughout your site.
Link from every important page to 2-6 closely related pages. Use descriptive anchor text that includes the target page’s keyword – avoid generic phrases like “click here” or “read more.” Keep important pages within three clicks of your homepage, because pages deeper in your site structure tend to receive less authority.
For AI search, internal linking helps establish topical authority. When your site has a clear cluster of interconnected content around a topic, AI platforms recognize that breadth of coverage as a sign of expertise. A site with a single blog post about “digital marketing” is less likely to be cited than a site with 30 interlinked posts covering specific aspects of digital marketing in depth.
Regularly audit your internal links to catch broken links and orphaned pages (pages with no internal links pointing to them). Both undermine your site’s crawlability and authority signals.
8. External Linking
Linking out to authoritative, relevant external sources signals to Google that your content is well-researched and exists within a broader context of quality information.
Link to reputable sources that support your claims – government websites, industry publications, peer-reviewed research and recognized authorities in your field. A best practice is to include at least 2-3 quality external links per page.
For AI search, outbound links carry even more weight. LLMs are looking for content that demonstrates it’s well-sourced and verifiable. A page that cites specific studies by name, links to primary research and references authoritative sources is significantly more likely to be cited by AI platforms than one that makes unsourced claims.
9. E-E-A-T: Experience, Expertise, Authoritativeness and Trustworthiness
E-E-A-T has gone from a quality guideline to a critical ranking factor. In 2026, it’s also one of the primary factors that AI platforms use to decide which sources to cite.
Here’s the reality: AI platforms are drowning in content. The web is flooded with AI-generated articles that say the same generic things. To stand out, your content needs to demonstrate that a real person with real expertise created it.
Experience. Show first-hand involvement with the topic. Use specific examples from your own work. Include original photos, screenshots, case studies and “I did this and here’s what happened” narratives. This is the hardest signal for AI-generated content to fake, which is exactly why Google and LLMs reward it.
Expertise. Attach named authors with verifiable credentials to your content. Create detailed author bio pages that include professional background, qualifications, links to LinkedIn profiles and other publications.
Authoritativeness. This comes from external recognition – backlinks from respected sites, mentions in media, guest appearances on industry podcasts and publications. You can’t self-declare authority – it has to be earned from others in your field.
Trustworthiness. Google has stated that trust is the most important E-E-A-T factor. Use HTTPS. Display clear business information (address, phone number, privacy policy). Cite your sources. Be transparent about who you are and what you do.
10. Schema Markup and Structured Data
Schema markup is a standardized vocabulary of code (typically in JSON-LD format) that you add to your pages to tell search engines and AI systems exactly what your content represents.
Optimizing your schema markup often has a strong positive impact on rankings, but in some cases it has very minimal impact or no impact at all. However, we still recommend doing it to have all your bases covered.
Here are some common schema types for on-page SEO:
Article schema – for blog posts and news stories. Includes properties for author, date published, date modified and headline.
FAQ schema – for pages with question-and-answer content. This is one of the lowest-effort, highest-impact schema types you can add.
HowTo schema – for step-by-step instructional content. Helps AI systems extract specific steps from your guides and tutorials.
LocalBusiness schema – for businesses with physical locations. Helpful for local SEO and AI platforms answering location-based queries.
Person schema – for author bio pages. Linking author credentials to your content through Person schema strengthens your E-E-A-T signals.
Product schema – for ecommerce pages. ChatGPT has confirmed it uses structured data to determine which products appear in its results.
Make sure your schema markup matches what’s actually visible on the page. AI systems check for consistency – if your Article schema says “Published: January 2026” but your page shows a different date, that’s a red flag. Validate your markup regularly using Google’s Rich Results Test.
11. Image Optimization
Images impact both user experience and search performance. They also create additional indexing opportunities through Google Images.
Compress images so they don’t slow down your page load speed. Use descriptive, keyword-rich file names (not “IMG_1234.jpg” but “kitchen-renovation-toronto-2026.jpg”). Write alt text that accurately describes the image in 125 characters or fewer – this is critical for accessibility and also helps Google and AI systems understand what the image depicts.
Use original images wherever possible. Stock photos don’t contribute to E-E-A-T signals the way original photos, screenshots and custom graphics do. If you’re writing about a process you’ve actually done, include photos that prove it. That first-hand evidence strengthens your “Experience” signal.
For AI search specifically, images with descriptive alt text and surrounding context help AI platforms understand your content more completely. If you include a chart or data visualization, make sure the key data points are also stated in the alt text – LLMs can’t read images.
12. Core Web Vitals and Page Speed
Google uses Core Web Vitals as part of its page experience signals. These metrics measure loading performance, interactivity and visual stability:
Largest Contentful Paint (LCP) measures how quickly the main content of your page loads. Aim for under 2.5 seconds.
Interaction to Next Paint (INP) measures how quickly your page responds to user interactions. Aim for under 200 milliseconds.
Cumulative Layout Shift (CLS) measures visual stability – whether elements on the page jump around as it loads. Aim for a score under 0.1.
These metrics matter for traditional SEO directly – they influence rankings and user experience. They matter for AI search indirectly – a slow, poorly performing page is less likely to rank well on Google, which means AI platforms are less likely to discover and cite it.
The most common fixes: compress and lazy-load images, minimize JavaScript and CSS, use a caching plugin and content delivery network (CDN), avoid layout-shifting elements above the fold and implement proper image dimensions so the browser doesn’t have to calculate them on the fly.
13. Mobile Optimization
Google uses mobile-first indexing, meaning it primarily evaluates the mobile version of your site for ranking purposes. If your mobile experience is poor, your rankings will suffer regardless of how good your desktop site looks.
Ensure your site is fully responsive. Test it on real mobile devices, not just by resizing your browser window. Make sure text is readable without zooming, buttons are large enough to tap and there are no horizontal scroll issues.
Over half of all web traffic comes from mobile devices, and a poor mobile experience directly hurts both your traditional rankings and your ability to be discovered by AI platforms.
14. Content Freshness and Update Strategy
Both Google and LLMs favour content that’s current.
Build a regular content update cycle into your workflow. Audit your top-performing pages quarterly. Look for pages with declining traffic, outdated statistics, broken links or information that’s no longer accurate. Refresh them with current data, new examples and updated recommendations.
Add visible “last updated” dates and/or include the current year in the page title, headings and body content. This signals freshness to both readers and AI platforms. LLMs have been shown to prioritize recently updated content when selecting sources to cite.
Don’t just change the date and call it an update. Make real improvements – add new sections, replace old data points, improve internal linking and incorporate recent developments. Search engines can detect superficial date changes without meaningful content improvements.
Putting It All Together
On-page SEO in 2026 isn’t a single tactic – it’s a system of interconnected signals that tell both traditional search engines and AI platforms that your content is relevant, authoritative and trustworthy.
Start with the fundamentals: strong title tags, clean URLs, clear heading structures and well-optimized content. Layer on the elements that strengthen AI visibility: answer-first writing, entity-rich language, schema markup and visible E-E-A-T signals. Then maintain it all with regular updates, internal link audits and performance monitoring.
The pages that win in both traditional search and AI citations share the same characteristics: they’re written by credible people, they answer questions directly, they cite their sources, they’re structured so machines can parse them easily and they’re kept up to date.
None of this is overly complicated. But it does require consistent execution across every page on your site. The businesses that treat on-page SEO as an ongoing priority are the ones that will hold positions that their competitors struggle to overtake.
PS: on-page SEO is only one part of the SEO equation – don’t forget about backlink acquisition strategies and technical SEO!
