How to Get Your Post-Production Studio Cited by ChatGPT and Google AI Overviews

We started paying attention to generative engine optimisation (GEO) the same week a client mentioned that their CMO had asked ChatGPT "which studios in Istanbul handle beverage CGI?" and got a short, confident answer that didn't include us. That's when the problem stops being abstract.

Two things about the current state of AI search worth anchoring. ChatGPT now drives roughly 87.4% of AI referral traffic. And the sources cited in AI answers overlap less than 20% with the top-ranked pages on Google (Superlines, eMarketer Q1 2026). In plain terms: being good at traditional SEO does not automatically make you visible in AI search. It's a related game with different rules.

This is what we've changed in our own content over the last six months, and what's working for studios like ours.

Why AI search doesn't cite your sales pages

The first honest thing to say: AI answer engines are not looking for marketing content. They are looking for answers. They reward specific, useful, verifiable writing; the kind that reads like a well-informed expert explaining something to a peer. They down-rank content that reads like a pitch.

The pages on most creative studio sites are pitches. "Our team of award-winning creatives brings your vision to life." A model can't extract a fact from that sentence. There is no information in it. When a brand asks "who handles AI imagery for beverage brands", the model passes over pages that say "we are passionate about storytelling" and cites pages that say "here's how we handle foil shaders on bottle renders".

The shift is uncomfortable because it exposes how much studio marketing copy has been filler. It's also a correction; a fair one.

What does AI-friendly content actually look like?

A few patterns we've seen work, ours and others:

• Question-form headings. "How does a CGI bottle render get quoted?" beats "Our Pricing Philosophy". Models match user questions to your page headings more reliably than they match to clever feature names.
• Named stats with sources. "Bespoke bottle renders run $800-$2,000 per angle" is the kind of fact an AI engine extracts and cites. Vague language doesn't get cited.
• Technical specificity. "We composite labels as UV-mapped textures with separate foil maps" reads as expertise. "We handle the details" reads as filler.
• Explicit process. Describing how you work, in ordered steps, with named stages, is both useful to readers and legible to models.
• Opinion with evidence. "Most 3D rendering companies quote per-SKU for work that should be shot-library based" is the kind of position an AI engine will quote; but only if you back it with reasoning.

What doesn't work

Equally useful to name what we've tried and seen fall flat.

• Listicles. "10 Reasons to Choose Our Studio" pages do not get cited. Models are trained to distrust that format. Listicles that do get cited are ones where the list is actually the natural structure (e.g., "6 file formats a CGI studio should deliver"), not a marketing frame bolted over a sales page.
• "Ultimate guides". Any page that has to call itself "ultimate", "complete", or "definitive" is almost always too shallow to earn the title. Models skim them and move on.
• Brochure-style case studies. A case study that reads as a self-congratulation piece ("we partnered with Brand X to deliver stunning results") gets ignored. A case study that reads as "here's the specific problem, here's what we tried first, here's what actually worked" gets cited.
• Pages with no date. Freshness signals matter. AI engines prefer content with clear publication or update dates.

Structure and metadata that helps

Some of this is unglamorous plumbing. Schema markup (Organization, Article, FAQPage) makes it easier for crawlers to understand what your page is. Clear publication and update dates in metadata. A sitemap that's actually current. Clean, semantic HTML with proper heading hierarchy. Alt text on images that describes the image, not the campaign.

None of these individually win citations; all of them remove friction from being picked up.

Where competitors still aren't looking

As of April 2026, the creative studio space is still under-invested in GEO relative to B2B SaaS and enterprise technology. The studios that publish answer-dense, opinionated, technically specific content on their own domains have a disproportionate opportunity; the competitive field for "which studio handles X" queries is thin. A single strong piece on a narrow topic, with named facts and a clear point of view, tends to outrank years of thin brochure content.

The other under-exploited area: answering practical buyer questions that brand managers actually ask. "How much should a CGI bottle render cost?" "How do I brief a studio for AI imagery?" "What's the difference between a render farm and a senior studio?" The studios that publish plain, well-researched answers to these questions start appearing in the AI results for buyers who are trying to figure this out.

What we changed at 35milimetre

Three specific things in the last six months. First, we shifted from writing about our services to writing about how the work actually gets done; process, pricing, file specs, specific problems we see in briefs. Second, we started publishing with named facts; prices, tool versions, dates, references to industry reports; rather than generalities. Third, we moved question-form headings to H2 across the blog, which is a small technical change that has had disproportionate impact.

We'll know in another few quarters how much this moves citations. What we can already see: the inbound conversations coming in are different. More specific. More technically informed. Often they start with "I read your piece on X and I have a brief on it."

Closing takeaway

GEO for creative studios is not a new discipline; it's the old discipline of writing useful things, with a sharper selection filter on the other end. The studios who will be cited by AI search in three years are the ones who spent 2026 publishing answer-dense, opinionated, technically specific content on their own sites. The ones who kept writing brochure copy will not be cited, which is a quiet catastrophe if your new business pipeline runs through AI recommendations.

If you're working on your own studio's content strategy and want to compare notes, we're always happy to talk.