Moz reframes SEO as relevance engineering for AI search growth

Google’s May 20, 2025 AI announcements made the shift concrete: AI Mode rolled out to all U.S. users, and AI Overviews expanded to more than 200 countries and territories in more than 40 languages. Moz’s relevance engineering framing gives agencies a sharper way to describe what clients actually need now: pages, sites, and brand signals that search engines and AI systems can interpret as credible, connected, and useful.

Relevance engineering changes the unit of work

Relevance engineering is the next layer on top of semantic SEO. The point is not to optimize isolated pages for one query, but to design a system of entities, topic coverage, and evidence that makes a site understandable to both humans and machine retrieval systems. AI Mode, AI Overviews, and LLM-based retrieval are changing how content gets surfaced, with more weight on passage-level optimization, topical depth, and Brand Authority.

For agencies, that language is useful because it moves the sales conversation away from tactical keyword work and toward a repeatable strategic service. A client does not buy “more content” when the real problem is that its site fails to establish what the brand knows, which entities it owns, and which intents it answers better than anyone else.

Google’s AI search shift made the timing obvious

Google’s May 20, 2025 AI announcements gave the market the practical reason to care. AI Mode uses query fan-out, breaking a question into subtopics and running multiple queries simultaneously.

Google’s own Search Central guidance points in the same direction. Its ranking systems are designed to prioritize helpful, reliable, people-first content, and the guidance asks whether a page provides original information, substantial coverage, and evidence of expertise. Internal links help people and Google make sense of a site, and structured data helps Google understand page content and eligibility for rich results.

What an agency has to change in delivery

Relevance engineering becomes a service line only if the agency changes how it works. Audits need to move beyond title tags and H1s and into entity coverage, intent mapping, trust signals, and information architecture. Content briefs need to name the primary entity, the supporting entities, the evidence to include, the stage of intent being served, and the adjacent pages that should link in and out.

The practical checklist looks like this:

• Map the core entities a client must own, then identify where the site already proves expertise and where it stays vague.

• Break topics into intent stages, from discovery and comparison to evaluation and conversion, so each page has one clear job.

• Build topic clusters around those intent stages, not around keyword variants alone.

• Add internal links that make the relationship between cornerstone pages, supporting pages, and commercial pages easy for both users and crawlers to follow.

• Use structured data where it clarifies page meaning, especially on product, service, FAQ, article, and organization pages.

• Rewrite briefs so they require original data, first-party experience, named examples, and explicit proof of expertise.

Why this sells better than standard SEO

The business case is straightforward. Clients facing AI visibility pressure, falling click-through rates, and generic content saturation are more willing to buy strategic work that improves the full relevance stack. “We will optimize your keywords” sounds interchangeable. “We will engineer your brand’s relevance across entities, evidence, and intent” sounds closer to the actual problem.

That also creates a better fit for agencies with vertical expertise. Relevance engineering is easier to package when the team already knows the language, trust sources, regulations, and buying patterns of a niche. In practice, that makes it a natural fit for agencies selling white-label services, authority-building retainers, and vertical-specific consulting programs that can be repeated across similar clients.

Mike King has become one of the clearest public examples of this shift, including under the label “chief relevance engineer.” By 2020, King and his team were experimenting with language models and building AI Overview-like tools.

How reporting has to change

If the offer changes, the dashboard has to change with it. Traditional reporting still matters, but it is not enough to track only rankings and sessions when the search experience is fragmenting across AI answers, passages, and branded discovery signals.

Agencies should add metrics that reflect the new buying environment:

• Branded visibility in AI surfaces and search results.

• Passage-level coverage for core questions and subtopics.

• Internal link discovery on priority pages.

• Presence of structured data on pages that need rich result eligibility.

• Directional traffic signals from non-head terms, not just exact-match rankings.

• Mindshare indicators in the client’s priority categories.