AI Citations Are the New Backlinks: How to Win Them

The rules of search visibility are being rewritten in real time — and the teams paying attention aren’t just optimising for rankings anymore. They’re asking a harder question: when an AI answers my customer’s query, does it mention my brand?

This isn’t a hypothetical concern for 2028. ChatGPT, Gemini, and Perplexity are already fielding product and service queries across Southeast Asia’s increasingly English-proficient urban professional class. If your brand isn’t in the answer, you don’t exist to that user — no matter where you rank on page one.

AI Share of Voice: The Metric Your Dashboard Is Missing

Most marketing teams are still measuring impressions and click-through rates from traditional SERPs. That’s fine for reporting to the CFO. It’s less useful for understanding where your brand actually lives in the minds of AI systems.

Semrush has introduced tooling to track AI share of voice — essentially, how frequently your brand appears in AI-generated answers relative to competitors for a defined set of queries. The methodology involves running target prompts through AI engines, then auditing which brands, URLs, and sources get surfaced. Think of it as a share-of-shelf metric, but the shelf is inside a language model.

For Southeast Asian brands, this matters acutely. A regional insurance company competing against global aggregators, or a homegrown e-commerce platform going up against Lazada and Shopee, needs to know whether AI is treating them as a credible source — or simply not registering them at all. The measurement comes first. Without a baseline, any optimisation effort is just noise.

Building AI Citations: Outreach Over Optimisation

Moz’s Charlie Marchant laid out a citation-building framework that should feel familiar to anyone who’s run a link-building programme — because structurally, it rhymes. The process starts with prompt research: identify the specific questions your audience is likely asking AI engines, then analyse which sources those engines are currently pulling from to construct answers.

What you find in that analysis is your target list. If Gemini consistently cites a handful of industry publications, analyst reports, or authoritative FAQ pages when answering queries in your category, those are the properties you need to appear on, be quoted in, or contribute to. Proactive outreach — getting your brand referenced in the right third-party content — is the primary lever, not on-page tweaks.

This is a meaningful shift for teams that have spent years in technical SEO mode. The citation graph that AI models navigate looks less like a keyword map and more like a reputation network. Authority is inferred from who talks about you, not just what you say about yourself.

---

What Google Is Quietly Signalling to SEOs

While the industry fixates on AI engines, Google itself is sending clear signals about where it wants the ecosystem to go — and some of those signals have direct implications for AI-readiness.

Search Engine Journal reported that Google is now warning against building separate markdown pages specifically engineered for AI crawlers. The logic tracks: creating AI-bait content that diverges from your primary web content is a variant of cloaking, and Google’s systems are likely to treat it accordingly. Mueller’s concurrent warnings about hidden link structures reinforce the broader principle — manipulation that’s invisible to users but present for machines tends to end badly.

The more interesting signal is Google’s new Search Console reporting for social posts and updated product markup guidance. Both suggest Google is expanding the structured data surface area it uses to understand and represent brands. For teams building toward AI citation authority, this is directly relevant: structured, machine-readable information about your brand and products is likely to inform how AI systems — including Google’s own AI Overviews — characterise you in generated answers.

The Southeast Asia-Specific Complication

Building AI citation authority in Southeast Asia isn’t a straight translation of Western playbooks. A few practical realities worth accounting for:

First, language fragmentation. AI engines are substantially better at synthesising English-language sources than Thai, Bahasa Indonesia, or Vietnamese content. Brands that have invested heavily in local-language content may find their AI visibility underperforms their actual market authority. Building English-language thought leadership — or ensuring key claims exist in English-accessible sources — is a legitimate tactical consideration, not a retreat from localisation.

Second, platform trust signals are different here. A Shopee seller with 50,000 reviews is authoritative in a Southeast Asian consumer context, but AI models trained primarily on Western web data may weight a single TechCrunch mention more heavily. Understanding which third-party sources AI models actually trust in your category — and pursuing presence there — requires regional prompt testing, not assumptions.

Third, mobile-first content structures. Long-form editorial content that AI models tend to cite well is often deprioritised in mobile-first publishing workflows. The answer isn’t to abandon mobile UX; it’s to ensure that substantive, citable content exists in a crawlable format alongside whatever mobile experience you’re serving users.

Key Takeaways

• Establish an AI share of voice baseline using tools like Semrush before attempting any GEO optimisation — you can’t improve what you haven’t measured.
• Treat AI citation building as a reputation programme, not an on-page exercise: the sources that cite you matter more than the content you publish about yourself.
• Resist the temptation to create separate AI-optimised content layers — Google’s guidance is clear, and divergent content strategies create long-term technical debt.

The brands that will own AI-cited authority in three years are almost certainly not the ones scrambling to reverse-engineer prompts in isolation. They’re the ones treating AI visibility as an extension of editorial credibility and third-party trust — which raises a harder question worth sitting with: if AI stripped away your paid media and your SERP rankings tomorrow, would your brand still be the answer?