Google's June spam update and AI citations: what Southeast Asian B2B brands should stop doing — and what actually builds visibility

Two things are true at the same time — and most B2B marketing teams in Southeast Asia are only acting on one of them.

First: buyers in Singapore, Malaysia, and Indonesia are shortlisting vendors inside ChatGPT, Perplexity, and Google AI Overview. If your company is absent from those answers, you lose deals before the first demo. We have documented that gap with audit evidence, not theory.

Second: on 24 June 2026, Google began rolling out a global spam update that enforces policies treating attempts to manipulate generative AI responses in Google Search as spam — the same category as link schemes and scaled content abuse.

Those two facts are not in conflict. But they create a trap for vendors who hear “you need AI visibility” and respond by buying their way into the answer.

This is the AI visibility trap — and Southeast Asian B2B brands are walking into it faster than most.

What actually changed

Google did not announce new spam rules on 24 June. It released an enforcement update — a SpamBrain upgrade applied globally and to all languages, with rollout expected to take a few days.

The policy shift that matters arrived earlier. In May 2026, Google expanded its spam documentation to state explicitly that spam includes attempting to manipulate generative AI responses in Google Search. That covers AI Overviews and AI Mode, not just traditional blue-link rankings.

Google’s own definition now reads, in part:

Spam refers to techniques used to deceive users or manipulate our Search systems into featuring content prominently, such as attempting to manipulate Search systems into ranking content highly or attempting to manipulate generative AI responses in Google Search.

The June update is the first major enforcement cycle to land after that expansion. Industry observers had already documented a pattern: self-promotional “best of” listicles surfacing as authoritative citations inside AI Overviews, often with thin substance and obvious commercial intent. Google has now formalised the response.

What this means in practice: if your visibility strategy depends on manufacturing mentions — paid placements on roundup sites, networks of comparison pages on expired domains, scaled AI-generated listicles seeded across third-party blogs — you are not building AI visibility. You are accumulating spam risk.

And spam demotion in Google Search does not stop at rankings. Sites that violate spam policies may rank lower or disappear from results entirely — which also removes them from the pool of sources Google uses when it generates AI answers.

The tactics now under active enforcement

Google’s spam catalogue did not get a separate “AI rulebook.” The same categories that applied to classic SEO spam now apply to AI surfaces. For B2B vendors evaluating GEO or AEO agency pitches, these are the patterns to treat as red flags:

None of this requires a buyer to report you. SpamBrain is automated. Recovery, when it comes, is measured in months — not days.

Why Southeast Asia is especially exposed

This is not because Google targets ASEAN markets specifically. The update is global. The exposure is behavioural.

Buyers here move fast. Cross-market audit work shows category shortlists forming in AI answers today — in English and in local languages. A Singapore HR director and a Jakarta HR manager do not get the same Perplexity shortlist. Language is a separate visibility profile, not a translation task.

Vendors here feel behind. Western incumbents and global review platforms already dominate many English-language AI answers. Local Series A and Series B companies respond with urgency — which makes “guaranteed AI citation” packages attractive.

The shortcut economy is loud. GEO playbooks imported from US and European markets often repackage old affiliate SEO: buy a mention on a listicle, spin up ten comparison pages, seed brand names across a blog network. In ASEAN, where category analyst coverage is thinner and affiliate roundup sites are plentiful, that playbook spread quickly.

The wrong metric creates false confidence. A vendor can appear in Google AI Overview on one query and be absent from ChatGPT and Perplexity on the same buyer prompt — Swingvy’s payroll audit showed exactly that pattern on 26 June 2026. Teams that celebrate a single AI Overview mention without auditing all engines — and without asking how that mention was earned — are optimising for a surface, not a strategy.

The companies most at risk are not the ones ignoring AI visibility. They are the ones panicking into manipulation because they correctly understand the stakes and incorrectly trust the shortcut.

Thin listicle vs substantive comparison

A fair question: if Google is cracking down on “best of” content, is all comparison content now toxic?

No. Google distinguishes content built to manipulate from content built to help buyers decide.

The pattern under enforcement looks like this:

• A page titled “Best HR software Singapore 2026” with a ranked table, affiliate links, and no mention of CPF, SDL, IRAS filing, or MOM record-keeping requirements
• The same template republished across .co, .com, and expired domains with swapped city names
• A brand that appears on the list because they paid for placement, not because an analyst evaluated the product

Substantive comparison content looks different:

• Local compliance requirements explained with links to primary sources (IRAS, MOM, BPJS, EPF — depending on market)
• Methodology stated — when data was retrieved, which buyer prompts were tested, what engines were queried
• Multiple vendors named with neutral framing and genuine selection criteria
• FAQ structured around how buyers actually research

Our Singapore payroll comparison follows the second model: IRAS controls, engine-by-engine citation differences, audit-backed shortlist data, and a methodology note. That is the kind of page AI engines cite because it helps a buyer — not because someone manufactured a mention.

If you are a B2B vendor, the test is simple: would this page still be worth publishing if AI Overviews did not exist? If the answer is no, Google now reads it as spam — and your buyers probably already did.

What to stop doing now

If any of the following is in your current or planned visibility programme, pause and audit before the next content sprint:

1. Paid listicle placements on sites whose primary purpose is ranking in AI answers or Google — especially networks that offer the same “best of” page across multiple domains
2. Agency contracts promising AI citations through mention seeding, third-party blog networks, or “recommendation poisoning” tactics
3. Scaled content programmes — AI-generated comparison pages across categories you do not actually serve, or across markets where you have no local product depth
4. Language expansion by translation — publishing Bahasa or Thai versions of thin English listicles without local compliance research
5. Celebrating AI Overview presence without source quality checks — if your only citations come from suspicious roundups, that presence is fragile

Link spam carries a specific warning in Google’s documentation: if you built rankings on manipulative links, removing the links does not automatically restore the benefit. The same logic applies to manufactured citation footprints. Unwind them deliberately; do not wait for a manual action email.

What actually builds visibility in Southeast Asia

The playbook is slower than a GEO shortcut. It is also the only one that survives enforcement — and the only one that works across ChatGPT, Perplexity, Claude, and Google at the same time.

1. Audit before you publish.

Run the exact prompts your buyers type — per country, per language. Record who gets cited, which sources each engine uses, and where you are absent. That baseline separates a visibility gap from a spam footprint. Request a free AI visibility audit or replicate the protocol from our ASEAN buyer research.

2. Publish primary substance, not mention bait.

Invest in content only your company can credibly produce: implementation guides grounded in local regulation, original research with stated methodology, customer evidence with specifics, product documentation structured for comparison queries. AI engines cite sources that reduce buyer uncertainty — not sources that exist to name your brand.

3. Earn third-party citations; do not buy them.

Comparison articles, review platforms, and analyst coverage matter because AI engines trust third parties for vendor shortlists. The goal is to be named in those sources because your product and market presence deserve it — through demos, integrations, case studies, and category participation — not because you purchased a row in a table.

4. Treat each market and language as a separate visibility profile.

Mekari Talenta appeared in English Perplexity but not Bahasa Perplexity in our Indonesia audit. Swingvy appeared in Malaysia ChatGPT but not Singapore ChatGPT. A single English content programme does not cover ASEAN. Neither does a translated listicle.

5. Track all engines, not just Google.

Google’s June update is Google’s enforcement event. Your buyers are not confined to one surface. A strategy that games AI Overviews while remaining absent from ChatGPT still loses deals — and now carries demotion risk on the Google side as well.

6. Clean up legacy content before you add new content.

If a previous agency built doorway pages, thin roundups, or scaled comparison networks on your behalf, inventory and remove them. Making your site compliant is a prerequisite; adding more content on top of a spam footprint does not fix the footprint.

The window vendors still have

Google’s May policy expansion and June enforcement are not a one-day event. Analysts expect citation-quality enforcement to ramp over subsequent quarters as SpamBrain learns new patterns.

That means two things for Southeast Asian B2B brands:

If you have been cutting corners, you still have time to audit and unwind — but the cost of delay rises with every enforcement cycle.

If you have been building substance, competitors who get filtered out of the citation pool make your earned mentions more valuable — in Google AI Overviews and in the conversational engines your buyers already use.

The visibility gap is real. The shortcut economy is a trap. Closing the gap correctly starts the same way it always should have: measure what AI engines say about your category today, then earn the citation with work buyers can trust.

If you want a baseline for your company, request a free AI visibility audit — or start with why Singapore B2B companies are invisible in AI search to see the audit methodology in practice.