AI SEO Strategy Is Not What It Was 12 Months Ago

AI SEO strategy in 2025 means something materially different from what it meant when most of the existing advice was written. The search engine is no longer just ranking pages, it is synthesising answers, and the content that wins is the content that gets cited inside those answers, not just the content that ranks in position one. If your SEO strategy was built around traditional ranking signals and keyword density, it is already behind.

This is not a crisis. It is a recalibration. The fundamentals of good content still matter, but the architecture around them needs to change. What follows is a commercially grounded look at how to build an AI SEO strategy that actually performs, drawn from what is working now rather than what worked three years ago.

What Has Actually Changed in AI SEO

I have spent time over the past year reviewing content performance across clients in a range of sectors, from professional services to e-commerce to B2B SaaS. The pattern is consistent: pages that were performing well in traditional search are not automatically performing well in AI-generated results. The two are not the same competition.

Traditional SEO rewarded pages that accumulated backlinks, matched keyword intent, and loaded quickly. Those signals still matter. But AI search adds a new layer: the model needs to trust your content enough to cite it. That means your content needs to be structurally clear, factually specific, and attributable to a credible source. Vague, padded content that ranked because it had the right anchor text is being passed over in favour of content that actually answers the question.

The Ahrefs team has covered this shift in detail in their AI SEO webinar series, and the consistent message is that topical authority and content structure are the two variables that matter most right now. That aligns with what I am seeing in practice.

If you want a broader view of how AI is reshaping the marketing toolkit beyond just search, the AI Marketing hub covers the full landscape, from content generation to performance measurement.

Why Entity Authority Is the New Domain Authority

Domain authority was always a proxy metric, not a real ranking signal. It was useful shorthand for how much trust a site had accumulated, but it was never the thing itself. AI search has made this distinction more visible, because the models are not just evaluating your domain, they are evaluating whether they know who you are.

Entity authority is about whether Google’s knowledge graph, and by extension its AI systems, can confidently connect your brand, your authors, and your content to a specific area of expertise. A site with moderate domain authority but clear entity signals, consistent authorship, structured data, and a coherent topical focus, will outperform a high-authority domain that publishes across fifty unrelated topics.

I saw a version of this years before AI search was a conversation. When I was building out content strategy for agency clients, the brands that owned a clear position in their category consistently outperformed those trying to be everything to everyone, even when the latter had bigger budgets. The principle is the same. Clarity of position is a commercial advantage, and now it is a technical one too.

Practical steps here include ensuring your authors have structured profiles with consistent bios across your site and third-party platforms, implementing schema markup that identifies your organisation and its subject matter, and building a content footprint that demonstrates genuine depth in your core topics rather than surface coverage of everything adjacent.

How to Build a Content Cluster That AI Systems Actually Use

Content clusters are not a new concept. The hub-and-spoke model has been recommended by SEO practitioners for years. What has changed is the reason it works. Previously, clusters worked primarily because of internal linking and crawl efficiency. Now they work because they signal topical completeness to AI systems that are trying to assess whether a source covers a subject thoroughly enough to be cited.

A well-constructed cluster for AI SEO looks like this: a central pillar page that provides a comprehensive, structured overview of the topic, surrounded by spoke pages that answer specific sub-questions with genuine depth. Each spoke should link back to the pillar, and the pillar should link forward to the spokes. The connections should be logical and contextual, not mechanical.

The mistake I see most often is brands building clusters that are wide but shallow. They publish twenty pages on a topic, but each page is 400 words of thin content that does not actually answer the question it promises to address. That is worse than publishing five pages of genuinely useful content. AI systems can identify thin content, and so can readers.

Moz has done useful work on how AI-generated and AI-optimised content performs in search, and the findings are worth reviewing before you scale any content programme. The short version is that quality signals, specificity, structure, and demonstrable expertise, consistently outperform volume.

The Role of AI Tools in Building the Strategy

There is an obvious irony in using AI tools to build an AI SEO strategy, and I want to address it directly. AI writing and research tools are genuinely useful for parts of the workflow. They are not a substitute for strategic thinking, and they are not a shortcut to authority.

Where AI tools add real value: identifying content gaps at scale, generating structural outlines for complex topics, drafting initial versions of factual content that a subject matter expert then reviews and enriches, and processing large keyword datasets to find clustering opportunities. Semrush has a useful breakdown of how AI copywriting tools fit into a content workflow, and it is honest about both the capabilities and the limitations.

Where AI tools do not add value: generating the specific, experience-based insights that make content worth citing. No AI tool can replicate the credibility that comes from a named expert who has actually done the work. That is precisely what AI search systems are looking for, and it is the one thing you cannot automate.

Early in my career, when I was running a small team and had almost no budget, I taught myself to code rather than accept that we could not build a proper website. The point is not the coding, the point is that constraints force clarity about what actually matters. AI tools are a constraint remover in some respects, but they do not remove the need to think clearly about what your content is for and who it is serving.

For a broader view of the AI tools landscape in marketing, the Ahrefs AI tools webinar covers a range of practical applications across content, research, and technical SEO.

Keyword Strategy in an AI-First Search Environment

Keyword research is not dead. It has changed shape. The shift is from targeting individual keywords to understanding the question ecosystem around a topic, and then building content that addresses that ecosystem comprehensively.

In practice, this means moving away from single-keyword targeting and towards intent mapping. For any given topic, there are typically four or five distinct user intents: definitional questions, comparison questions, how-to questions, validation questions (is this the right approach?), and decision questions. A strong AI SEO strategy maps content to all of these intents, not just the high-volume head terms.

I spent years managing paid search campaigns across dozens of industries, and the discipline that made those campaigns work was the same one that makes content strategy work: understanding what the user is actually trying to accomplish, not just what words they typed. When I ran a paid search campaign for a music festival at lastminute.com and saw six figures of revenue come in within a day, it was not because we had the highest bids. It was because we had matched the right message to the right intent at the right moment. Content strategy is slower, but the underlying logic is identical.

The Semrush AI SEO assistant is worth exploring for intent-based keyword clustering, particularly if you are managing a large content programme and need to process significant volumes of keyword data efficiently.

Technical Foundations That Support AI Search Visibility

Technical SEO has not become less important in an AI search environment. If anything, it has become more important, because the signals that help AI systems parse and understand your content are largely technical.

Schema markup is the most direct lever. Structured data tells search engines exactly what type of content a page contains, who wrote it, what organisation published it, and how it relates to other content on the site. FAQ schema, Article schema, and Person schema are the three most relevant for content-heavy sites. If you are not using them, you are making the AI system work harder to understand your content, and it will often default to sources that make the job easier.

Page speed and Core Web Vitals remain relevant, not because they directly influence AI citation decisions, but because they influence whether Googlebot crawls and indexes your content efficiently. Content that is not indexed cannot be cited. This sounds obvious, but I have audited sites where significant portions of the content library were either not indexed or indexed with errors that prevented proper understanding of the page structure.

Internal linking architecture matters more than most brands realise. A well-linked site helps AI systems understand the relationships between your pages and the relative authority of different content within your domain. Orphaned pages, pages with no internal links pointing to them, are effectively invisible to the content graph that AI systems build when evaluating a site.

Measuring AI SEO Performance Without Losing Your Mind

Measurement in AI search is genuinely harder than in traditional search, and anyone telling you otherwise is either selling something or has not looked at the data closely. AI Overviews and other AI-generated results do not always pass click data through in the same way that traditional blue links do. Attribution is messier.

The honest approach is to track a combination of signals rather than looking for a single clean metric. Branded search volume is a useful proxy for awareness and citation, because when AI systems cite your content, some users will subsequently search for your brand directly. Impression data in Google Search Console, particularly for queries where you know AI Overviews are triggering, gives you a sense of visibility even when click-through rates are low. Direct traffic trends can indicate whether AI-driven awareness is translating into site visits through channels that bypass traditional search.

I have always been sceptical of measurement frameworks that prioritise clean reporting over honest approximation. When I was managing large media budgets, the temptation was always to over-index on the metrics that were easy to track rather than the ones that actually mattered. AI SEO is forcing the same discipline: you need to be honest about what you can and cannot measure, and build your reporting around directional signals rather than false precision.

Moz has published useful guidance on integrating AI tools into a content and measurement workflow that is worth reading if you are trying to build a practical reporting framework rather than a theoretical one.

The Competitive Reality: Most Brands Are Behind

Here is the commercially useful part of this picture. Most brands have not yet adapted their content strategy for AI search. They are still producing content to the same brief they were using three years ago, optimising for the same metrics, and wondering why organic performance is flattening despite consistent output.

This is a window. The brands that restructure their content architecture now, build genuine entity authority, and produce content with the structural clarity that AI systems need, will establish a position that becomes harder to displace over time. Authority compounds. The brands that wait until AI search is fully mainstream will find that the ground has already been taken.

I have seen this pattern play out repeatedly across the industries I have worked in. The businesses that moved early on paid search when it was still relatively cheap and poorly understood built advantages that took competitors years to close. The mechanism in AI SEO is different, but the commercial logic is the same. Early movers who do the work properly build compounding advantages. Late movers pay a premium to catch up, and often never fully close the gap.

The practical implication is not to move recklessly. It is to prioritise the foundational work: entity signals, content cluster architecture, schema implementation, and genuine topical depth. These are not speculative bets on where AI search is going. They are investments in the quality signals that have always mattered, now made more visible by AI systems that reward them more directly.

There is more on how AI is reshaping the broader marketing function, from content to media to measurement, in the AI Marketing section of The Marketing Juice. If AI SEO is a priority, the surrounding context matters.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions