FAQPage schema does two jobs at once. It tells Google AI Overviews and Google AI Mode that your page has structured question and answer content, and it gives the language models a clean, machine-readable extract to quote from. The pages that get pulled into a zero-click AI answer almost always have FAQPage markup applied to a real FAQ section, not to a marketing tagline labelled as a question.
Build the FAQ section first as plain visible content, with each question as a sub-heading and each answer as a short paragraph of 40 to 80 words. Then add the FAQPage JSON-LD that mirrors that visible content exactly, and test it in Google's Rich Results tester. Fresheather's services hub uses this pattern and links out to the relevant blog post for the long-form answer behind each question.
Perplexity weights freshness heavily, and Google AI Overviews quietly down-rank content with old or missing publication dates. Adding a clearly visible "updated 2026" line at the top of every cornerstone page is one of the lowest-effort signals you can send, and it materially changes which pages get sampled. The rule is to refresh the page meaningfully rather than only bumping the date, then publish a short note about what changed so the new date is honest. Pair it with a refresh of any year-specific stats inside the post so the numbers and the date agree.
Listicles such as "top creative agencies London", "best UK production companies" and similar category roundups are disproportionately cited by ChatGPT and Perplexity when answering shortlisting queries. The fastest path in is to pitch the writers of the roundups that already rank for your category, with a short, specific email that includes a sentence the writer can lift directly. Once you are inside two or three of those lists, your brand starts appearing in AI answers to shortlisting questions without any further work. Fresheather's brand activations page is structured to be liftable by exactly this kind of writer.
YouTube is the most-cited single domain in Google AI Overviews overall, and ChatGPT references YouTube transcripts heavily for how-to and definitional queries. One strong eight-minute explainer that answers a core category question, with a clean transcript and a descriptive title, can out-perform a year of blog content for AI citation volume. Treat YouTube as long-form FAQ infrastructure rather than as a marketing channel, and the work compounds.
Google AI Mode samples Reddit heavily. The brands that surface in Reddit-fed answers are the ones that have genuine, unsponsored mentions in category subreddits, usually from users or founders who actually participate in those threads. The work is not a posting schedule; it is consistent presence over six to nine months, answering questions in the threads that are already ranking for your category queries. The patient version of this tactic compounds faster than the impatient version, because moderators and Google both reward continuity.
LinkedIn is the most-cited domain for professional queries in Google AI Overviews. Consistent founder content on a single, well-defined topic raises brand-level citation rates because the language models build a statistical picture of who is authoritative on which subject. Two posts a week on one topic for three months is more useful than daily posts across five topics, because depth on a narrow subject is what the citation graph actually rewards.
Podcasts and trade press matter because the transcripts and the editorial coverage create co-mentions on high-trust domains. Pitch a small number of well-chosen shows and publications with a specific angle and a usable hook, rather than running a broad PR push. Each placement is a long-lived signal that the language models will re-encounter every time they crawl that domain, and a small handful of strong placements outperforms a large handful of weak ones. Fresheather's services hub lists the work types we use to anchor those pitches.
Original numbers, especially UK-specific ones, get cited because the language models prefer cite-ready data over commentary. A short, well-sourced piece of original research, published with clear methodology, will often outperform a long thought-leadership essay for AI citation volume. The trick is to make the headline numbers easy to quote in a single sentence, with the source visibly attributed to your brand and the methodology one click away.
The work only compounds if you measure it. Once a quarter, run the same set of 10 to 20 priority queries through ChatGPT, Perplexity, Google AI Overviews, and Google AI Mode, and log which domains get cited and which do not. The point is not the snapshot but the trend. The first audit gives you a baseline, the second tells you which tactics are working, and the third lets you redeploy effort away from anything flat.
ChatGPT leans on YouTube, Wikipedia, and a long tail of editorially-credible niche publications, with a clear bias toward content that answers the question in its first paragraph. Perplexity surfaces four to six sources per answer, weights freshness, and over-indexes on credible trade publications and academic sources. Google AI Overviews appear on roughly half of tracked UK queries and pull citations from a wider mix that leans on YouTube, Wikipedia, Reddit, and LinkedIn for category-level questions. Google AI Mode disproportionately samples Reddit and LinkedIn, especially for shortlisting and decision-support queries. The reason a brand needs all three workstreams together is that no single engine relies on a single source type, and a brand that is only visible on one of them is partially invisible across the others.
Expect the first citations within four to eight weeks of running the work in parallel: an answer-first cornerstone page, FAQ schema, and at least one strong off-site mention on YouTube, Reddit, or a credible trade publication. Single-channel work, whether that is only blog posts or only outreach, usually takes longer because the language models need several co-mentions on different domains before they treat your brand as a trusted source.
Yes. Perplexity links to four to six sources per answer, weights freshness heavily, and over-indexes on credible niche publications and academic content. Google AI Overviews sample a wider mix that leans on YouTube, Wikipedia, Reddit, and LinkedIn for category-level questions. The overlap is high, but the off-site signal mix that wins on Perplexity is not identical to the mix that wins on Google AI Overviews, which is why the audit in tactic 10 separates the engines rather than averaging them.
Yes, and arguably more easily than competing for blue-link rankings. Only 38 percent of Google AI Overview citations come from the top ten organic results, so a brand that is invisible on classic SEO can still surface inside AI answers if its content is answer-first, its schema is clean, and it has a small, consistent pattern of co-mentions on credible off-site domains.
Rewriting the first 60 to 100 words of every cornerstone page so they answer the page's core question in plain prose, with no preamble. That single change is the most common reason a page gets pulled into a ChatGPT, Perplexity, or Google AI Overview answer. The schema, the off-site work, and the date refresh stack on top of it, but the rewrite is what unlocks the rest.
Yes. Fresheather covers the cornerstone-content and earned-media layer of AI search optimisation and works in parallel with a brand's existing technical SEO and development teams. The split that tends to work best is: in-house SEO owns crawl, schema rollout, and Search Console; Fresheather owns the answer-first content, FAQ structure, and the off-site co-mention layer that drives citations on ChatGPT, Perplexity, and Google AI Overviews. The services hub covers the cornerstone-content remit in full.
For the deeper strategy primer behind these tactics, read the parent guide on how to get your brand cited by AI search.