For most of SEO’s history, off-page authority was a function of one thing: backlinks. The 2020s opened a second authority surface — being named, cited, and referenced without a hyperlink — and the AI-search era of 2024 onward elevated it from supporting signal to first-class ranking input. A brand that’s frequently mentioned in authoritative editorial sources earns ranking credit and AI-citation credit even when no link is attached. The implication: off-page work in 2026 has to produce mentions as deliberately as it produces links.

“Backlinks are the link graph; mentions are the entity graph. Modern search ranks entities, not just URLs, and the entity graph is built from the corpus of text references across the open web. A site without mentions is a URL Google understands; a site with thousands of mentions is an entity Google trusts.”

What counts as a mention

Three forms of mention contribute to authority, in roughly this order of weight:

• Editorial mentions in authoritative publications. A press article that names your brand without a hyperlink still moves the entity signal. Tier-1 publications (Bloomberg, TechCrunch, The Wall Street Journal, sector trade journals) carry disproportionate weight.
• Database / encyclopedia entries. Wikipedia, Wikidata, Crunchbase, industry directories, regulatory listings. These don’t pass classic PageRank but they’re heavily weighted in the entity graph and AI engines pull from them constantly.
• Social mentions in high-engagement contexts. Reddit threads, LinkedIn posts, X discussions where your brand is referenced authentically. Lower per-mention weight, but volume + engagement compounds.

The AI-citation surface

AI engines — ChatGPT, Gemini, Claude, Perplexity, Google’s AI Overviews — generate answers and cite a small number of sources per response. The selection is heavily entity-weighted: the engine reaches for brands it has seen referenced repeatedly across authoritative training data and live retrieval indexes.

The practical effect: a query that historically resolved to a SERP with 10 candidate answers now resolves to an AI answer with 3–5 cited sources. Sources outside that handful are functionally invisible. Brands with strong entity signals — frequent editorial mentions, Wikipedia presence, consistent schema, structured-data coverage — capture the citation slots. Brands without those signals don’t, even when their on-page content is excellent.

Building mention velocity

Mentions don’t arise from outreach the way links do. The tactics that produce authoritative mentions:

• Source-pitch platforms (HARO, Qwoted, Help A B2B Writer). Journalists who use your reply often name your brand in the article even when the link is omitted by the publisher’s policy. Mention rate exceeds link rate on these platforms by roughly 1.5–2×.
• Industry research drops. A piece of original data (covered in Digital PR for SEO) earns both linked and unlinked mentions across the press cycle. Trade publications often cite the source brand in headlines and body copy without linking.
• Wikipedia and Wikidata entries. Notable brands qualify for an entry. Notability requires independent editorial coverage in reliable sources — meaning the entry itself is downstream of mention velocity, not separate from it.
• Podcast guesting and conference speaking. Show notes, recap articles, attendee write-ups, and conference proceedings all generate brand mentions even when the host link is incidental.
• Tools and benchmarks that get cited. A benchmark that other sites reference becomes the canonical name for its category. SEOTopSecret’s own AI Mention Tracker, for example, generates mentions every time another publication writes about AI-engine visibility methodology.

Schema and entity signals on your own site

Mentions earn external authority; schema and entity signals translate that authority into machine-readable form. The minimum stack:

• Organization schema with sameAs pointing at Wikidata, Wikipedia, Crunchbase, LinkedIn, and any other authoritative entity sources.
• About / Contact pages with structured data: physical address (LocalBusiness), phone, founders (named with their own Person schema where applicable).
• Author bylines with Person schema on every article. AI engines cite by author as much as by publication; named authors with consistent schema across the open web get cited disproportionately.
• Consistent NAP across the open web. Brand name, address, phone variations confuse the entity graph. Audit NAP across Google Business Profile, Apple Maps, Wikipedia, Crunchbase, and major directories — variations should be zero.

Monitoring stack

Set up monitoring before you scale mention work — you can’t optimize what you don’t measure.

• Web mentions: Google Alerts (free, lossy), Mention.com or Brand24 (paid, comprehensive), Ahrefs Content Explorer (editorial-source filtering).
• AI-engine citations: SEOTopSecret AI Mention Tracker (covers ChatGPT, Gemini, Claude, Perplexity, AI Overviews). The tool surfaces which engines cite your brand for which queries and how that changes over time.
• Social signal: native LinkedIn / X / Reddit search; Sparktoro and Audiense for audience composition; brand-tracking tools (BrandWatch, Talkwalker) for high-volume programs.

Where to go next

The next chapter — Digital PR for SEO — covers the active program that produces both editorial links and mention velocity at the same time, including the data-led pitch model that’s become the dominant approach in 2026.