The modern SEO guide for 2026 (and what AI changes)

For most of its history, SEO has meant "rank higher on Google." That definition still describes the largest share of the work, but it is incomplete. Today, organic search results are layered: AI Overviews sit above the blue links on Google for a growing share of queries; featured snippets, knowledge panels, and shopping carousels eat real estate; Discover and News surface content outside the classic search box; and a parallel set of AI-first engines — ChatGPT Search, Perplexity, Gemini, Claude, Copilot — answer the same questions from the same underlying web.

What that means in practice is that the SEO discipline now has two interlocking jobs. The first is the classic one: rank well in organic results across the surfaces that already drive your traffic. The second is making sure your pages are structured, sourced, and trusted enough to be cited by the AI surfaces that are reshaping the top of the funnel. Both jobs share most of their foundations — the same content, the same site architecture, the same authority signals. The mix of effort is what shifts.

Every SEO programme — whether it is a one-person founder doing it themselves or a 40-person agency running it for hundreds of clients — rests on three pillars. They have not changed in twenty years, and they will not change in the next five.

The pillars are interdependent: technical hygiene lets the engine see what you have; content gives it something worth ranking; off-page authority tells it your site is a credible source for that content. Weakness in any one pillar caps the return on investment in the others. A perfectly written page on an unindexable URL ranks nowhere; a well-linked site full of thin content stalls; a structurally pristine site with no inbound links struggles to break through in a competitive category.

Most failing SEO programmes are not failing for lack of ideas — they are failing because one pillar is starved. Diagnose the weakest pillar first; the others will not pay off until it is fixed.

The single biggest shift in modern SEO is not technical — it is the move from keyword targeting to intent targeting. Google's ranking systems read the query for what the searcher actually wants and reward pages that satisfy that want, not pages that match the keyword most literally.

The practical version of this is the topic cluster: a hub page that defines a topic broadly, surrounded by spoke pages that go deep on subtopics, all internally linked. The hub captures the high-level query ("crm for accountants"); the spokes capture the specific intents around it ("crm for small accounting firms", "best crm for cpa workflow", "crm with quickbooks integration"). Together they convince the engine that you have topical authority on the cluster — and topical authority is what wins the high-volume head terms.

Three rules for clusters that actually work:

• One intent per page. Trying to satisfy three different intents on one page satisfies none of them. Each spoke gets a single search intent and serves it completely.
• Internal links matter more than people think. The hub links down to every spoke; every spoke links back up to the hub and across to its closest siblings. Most sites underinvest here.
• Build for completeness, not for volume. A cluster with 8 strong spokes covering the real subtopics beats a cluster with 30 thin spokes. Density of useful coverage is what the engine rewards.

Technical SEO is the work of making sure search engines can find, crawl, render, and understand your pages. It is the least glamorous pillar and the one that most often holds programmes back without anyone noticing.

The non-negotiables in 2026:

• Crawlability — robots.txt opens what should be open, an XML sitemap exists and is current, internal linking creates a path to every page that matters.
• Indexation — pages you want ranked are indexable; pages you do not (faceted-search permutations, staging, internal search results) are excluded with canonicals or noindex.
• Render parity — what Googlebot renders matches what a user sees. SPAs need server-side or static rendering for any page that has to rank.
• Core Web Vitals — LCP, CLS, INP within the "good" thresholds on real-user data. Slow pages get demoted; janky pages get demoted further.
• Schema markup — JSON-LD that confirms what the page is (Article, Product, Service, FAQPage). It does not directly cause rankings but it stops the engine from second-guessing.
• Mobile-first — the mobile rendering is the rendering. A great desktop page that breaks on a phone is treated as a broken page.

The trap teams fall into is treating technical SEO as a one-time project. It is a continuous one — every release ships new templates, every redesign shifts URL patterns, every CMS update changes how pages render. A monthly crawl with prioritised fixes is the minimum cadence for a site of any complexity.

Google's framework for evaluating content quality is summarised in the acronym E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness. The acronym is a useful shortcut, but the underlying principle is older and simpler: satisfy the searcher's intent more completely and more credibly than anything else competing for the query.

What separates the pages that consistently win from the pages that quietly lose:

• First-hand experience. Original data, original analysis, named authors with real credentials, photos and videos of you doing the thing — all read as experience signals.
• Depth without padding. The page covers everything the searcher needs without filler. Cut the throat-clearing introductions and the SEO-checkbox FAQs that pad word count without adding value.
• Clear sourcing. Claims are cited, statistics are dated, external sources are linked. The page reads like a credible reference instead of a content-marketing artifact.
• Author attribution and bio. Real authors with consistent bylines across the site. Generic "Editorial Team" attributions are a soft demerit on competitive YMYL topics.
• Maintenance. A "last updated" timestamp that is actually updated when the page is reviewed. Stale pages on time-sensitive topics get filtered out of the running.

Off-page SEO is the work of being recognised, by the wider web, as a credible source. Links are still the dominant signal — Google has said so directly, every credible study confirms it — but the modern off-page picture is broader than backlinks alone.

• Editorial links from credible neighbours. One link from a publication or peer that your category trusts beats a hundred links from a syndication farm. Earn the right ones; the wrong ones are at best ignored and at worst penalised.
• Brand mentions, even without a link. When credible sites mention you by name in context, the engine treats it as a soft authority signal. A consistent press footprint helps even when the journalists do not link.
• Knowledge-graph presence. Wikipedia, Wikidata, Crunchbase, LinkedIn, Google Business Profile — the structured-data layer the engines reconcile against. Get your brand entity right in those places and the rest gets easier.
• Reviews and social proof. Third-party reviews on G2, Capterra, Trustpilot, and the relevant industry sites accumulate as authority for the brand entity, not just the listing pages.

The fastest way to lose ground on the authority pillar is to ignore it. The fastest way to win is to build a beat — PR, partnerships, original research, executive presence on the channels your buyers read — and stay consistent with it for years.

The metrics that matter, in roughly the order an executive should care about them:

1. Organic revenue and pipeline. The number every SEO programme has to defend. Tie keyword and page work back to conversions, not just clicks.
2. Organic sessions, segmented. Total is a vanity number; segmented by intent stage (top of funnel vs. bottom of funnel), by page type, and by query type is where the signal lives.
3. Indexed pages vs. submitted. The single best early-warning indicator of a technical problem. A sudden gap between submitted and indexed means something is broken; fix it before traffic moves.
4. Rankings on your priority queries. Rank tracking is a leading indicator of traffic. Track the queries that matter; ignore the ones that do not.
5. Share of voice in your category. Your visibility relative to a defined competitive set. Easier to defend in a board meeting than absolute rank movement.
6. AI citation rate. The newest of the metrics, and increasingly the leading indicator for top-of-funnel demand. Covered in depth in the GEO Guide.

The honest answer to "how does AI change SEO" is: it changes the surface, not the foundation. Quality content, sound technical hygiene, and credible authority are still the inputs. What changes is the way those inputs are rewarded — and the new ways they can be wasted.

The two shifts that matter most:

• The lede matters more than ever. AI engines extract from the top of the page. A great page with a slow introduction loses to a good page with a sharp one. Tighten the first 100 words on every important page; you are writing for retrieval, not just engagement.
• Structure beats prose. Tables, numbered lists, named sections, and explicit definitions are easier for both classic search snippets and AI synthesis to lift. Nuance still belongs in the page — give the model a structured spine to find first.

The other shift, less discussed but increasingly visible: AI engines weight host authority similarly to how classic search does. The brand-building work the off-page pillar has always required pays off twice now — once for blue-link rankings, once for AI citations on the same domain.

The full breakdown of how GEO and SEO interact lives at GEO vs SEO: the comparison guide.

• Optimising one pillar in isolation. Content quality without technical hygiene wastes the content; technical perfection without content has nothing to rank.
• Chasing keyword volume instead of intent. High-volume head terms with vague intent rarely convert; specific long-tail with clear intent does most of the revenue work.
• Building thin clusters. Thirty spoke pages that each say a little is worse than eight that each say a lot. Density wins.
• Treating SEO as a project, not a programme. Every site ships breaking changes constantly. Without a monthly cadence of audit, content, and authority work, position decays.
• Ignoring AI search. The top of the funnel is moving into AI answers faster than most teams expect. A GEO programme alongside SEO is no longer optional in competitive categories.

A modern SEO programme needs five capabilities, and the teams that win consolidate them rather than stitching a dozen tools together.

1. Crawl and audit that runs on a cadence and prioritises issues by severity and effort.
2. Rank tracking across the queries that matter to your business, segmented by intent.
3. Content intelligence that finds gaps, surfaces striking-distance keywords, and prevents cannibalisation.
4. AI visibility tracking for citation rate, position, and competitor share of voice across the major AI engines.
5. Unified reporting that combines GSC, GA4, rank, and AI signals in one view your team and your stakeholders can both use.

Aergos was built to be that single platform. Technical SEO handles the audit pillar, Content Intelligence handles the strategy pillar, AI Visibility handles the new pillar, and Analytics & Reporting brings them together.