Technical SEO in 2026: Core Web Vitals, schema markup and AI search

Technical SEO in 2026 is no longer just covering keywords: Google rewards real-user performance (Core Web Vitals), understands pages better when they ship structured data, and the new AI-driven search experiences — SGE, Perplexity, ChatGPT Search — cite content based on very specific signals. This guide focuses on the practices that move the needle the most.

Core Web Vitals in 2026

Since March 2024, INP (Interaction to Next Paint) replaced FID. The three current metrics:

Metric	What it measures	"Good" threshold
LCP (Largest Contentful Paint)	Time to paint the largest visible element	≤ 2.5 s
INP (Interaction to Next Paint)	Latency between an interaction and the next paint	≤ 200 ms
CLS (Cumulative Layout Shift)	Visual stability of the content	≤ 0.1

Three levers that move these metrics on most sites:

• Image optimization: AVIF/WebP, loading="lazy", correct sizes, priority on the LCP element.
• Cut blocking JS: minimize hydration, use next/dynamic with ssr: false only on truly interactive components, split into islands rather than hydrating the whole page.
• Fonts: next/font with display: 'swap', preload the primary weight.

For real diagnosis, measure your site with CrUX data (not just lab Lighthouse) using our website speed test or Google Search Console.

Schema markup: the invisible SEO that actually counts

Schema.org remains the language Google uses to understand what's on your page. For an agency or local business, the minimum effective set:

1. Organization (or LocalBusiness / ProfessionalService if you have a physical address): name, logo, address, phone, hours, sameAs to social profiles.
2. WebSite with a SearchAction so Google can show the SERP sitelinks search box.
3. BreadcrumbList on inner pages — improves the snippet with the navigation path.
4. Service for each offering: name, description, serviceType, areaServed.
5. FAQPage on pages with frequently asked questions — can show up as rich results.
6. Article or BlogPosting on posts: author, date, image.

Tip: centralize schema in a single helper and reference entities with @id to link them (provider: { '@id': '/#organization' }). Avoids duplication and keeps the graph coherent.

GEO and AEO: optimizing for generative search

AI-powered engines (Google SGE, Perplexity, ChatGPT Search, Copilot) cite specific pages when answering. Signals that increase the likelihood of being cited:

• Clear, question-shaped headings: "What is X?", "How to Y", "Difference between A and B".
• Executive summaries at the start — models tend to extract the first synthesizable answer.
• Citations to sources and verifiable data; AI prefers content that looks authoritative.
• Semantic structure: headings, lists, tables. Easier to extract than dense paragraphs.
• Complete schema — models lean on structured data when it exists.

GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) aren't a radical new technique: they're SEO done well, with a focus on concise answers and structured data.

Tools we use in every technical SEO audit

• PageSpeed Insights and CrUX dashboard for real Core Web Vitals.
• Google Search Console: coverage, Core Web Vitals by origin, usability issues.
• Schema Markup Validator (validator.schema.org) and Rich Results Test for structured data.
• Screaming Frog or Sitebulb for link, redirect, and hreflang auditing.
• Server logs to understand what Googlebot actually crawls.

30-day action plan

1. Measure LCP/INP/CLS with real-user data (CrUX or our speed test).
2. Audit your schema with the official validator. If you're missing Organization, WebSite or BreadcrumbList, those are quick wins.
3. Review the titles of your top 10 traffic pages — do they answer a specific question?
4. Add a summary or "Conclusion" section to long posts. Improves AI extraction.
5. If the audit reveals serious performance issues, get in touch — performance is one of our pillars.

Conclusion

SEO in 2026 rewards real speed, semantic clarity, and structure. Pages that load fast, declare what they are with proper schema, and answer concrete questions through clear headings, are the ones that appear in both classic results and the generative answers from Google, Perplexity and ChatGPT.