AI search engines (ChatGPT, Perplexity, Google AI Overviews) — what Thai businesses need to do

By early 2026, Google's AI Overviews appear on ~28% of search queries. ChatGPT, Claude, Gemini, and Perplexity collectively handle a meaningful chunk of research queries that used to go to Google. "AI search" is real. Most advice about how to optimise for it is hype.

Here's what we actually see working in our client engagements.

All the major AI engines (Perplexity, ChatGPT search, Claude with search, Google AI Overviews, Bing Chat) work roughly the same way underneath: they run a traditional search, retrieve the top 5-20 sources, then synthesise an answer that cites those sources.

That means traditional SEO still matters. The pages that get cited in AI answers are the same pages that rank in normal search results, with three additional weights:

1. E-E-A-T signals: author credentials, dates, real organisation behind the content 2. Schema markup completeness: structured data lets the AI parse the content efficiently 3. Factual specificity: AI engines preferentially cite pages with concrete numbers, dates, and named entities over vague prose

### "AI SEO" agencies promising LLM-specific optimisation

Most agencies pitching "AI SEO" are selling the same SEO services they sold last year with new vocabulary. There's no separate optimisation channel for ChatGPT versus Google — they both pull from the open web.

Treat "AI SEO" as a marketing label, not a technical category. Anyone telling you they have a special way to rank in ChatGPT is overselling.

### Stuffing pages with AI-generated content

Google's helpful-content updates have specifically penalised AI-generated content that doesn't add value. Sites that ramped up AI publishing in 2024-2025 took 30-60% traffic drops in 2026 algorithm updates.

The useful pattern: AI-assisted writing (human edits, fact-checks, adds original perspective) — not AI-published.

### Targeting "AI keywords"

There's no special keyword research category for AI search. The queries that get AI Overview answers in Google overlap heavily with the queries that have featured snippets. Same keyword research approach.

### 1. Author bylines + bios

Google's AI Overviews and ChatGPT search both visibly prefer cited sources with named authors who have bios + credentials. A blog post by "Kanoktip Lergdee — CEO, Backlink Hut, 15 years in Thai SEO" gets cited more often than the same content with no byline.

Practical: every blog post and service page should have an author byline. The bio should link to a full /authors/{name} page with credentials, expertise areas, and other publications.

### 2. Updated-date and original-publish-date

AI engines prefer recent content for queries about evolving topics ("best WordPress hosting 2026", "Google Ads changes"). Show the publish date AND last-updated date visibly. Mark old posts as "last reviewed" with the current date when you re-audit them.

Google reads the schema `datePublished` and `dateModified` fields. AI engines weight these heavily for time-sensitive topics.

### 3. Schema markup completeness

The sites getting cited most in AI Overviews have rich schema:

• Article schema with author, publisher, datePublished, dateModified
• FAQPage schema on every page with Q&A
• HowTo schema on step-by-step content
• Organization schema with knowsAbout listing expertise areas
• BreadcrumbList schema for content hierarchy

Most Thai SMB sites have zero schema. Even basic Article + FAQ + BreadcrumbList on every page dramatically improves AI citation rate.

### 4. Specific entities, numbers, and dates

Vague content ("SEO usually takes some time") rarely gets cited. Specific content ("SEO takes 60-90 days for long-tail keywords, 6-12 months for competitive head terms in Bangkok B2B niches") gets cited constantly.

The pattern: name specific places, name specific numbers, give specific date ranges, cite specific examples. Concrete > vague.

### 5. FAQ format throughout

Most AI Overview citations come from FAQ-style content because AI engines find Q&A pairs easy to extract and cite. Add FAQs to every service and location page, not just a dedicated /faq page.

### 6. Direct, declarative prose

AI engines extract sentences as citations. Sentences with clear declarative structure get cited more than sentences buried in long flowing paragraphs.

Pattern: lead with the fact, then justify. "Map Pack accounts for 40-60% of local clicks. Most Thai SMBs lose this because of NAP inconsistency." Each sentence stands alone.

Thai-language queries in AI engines have specific dynamics in 2026:

• ChatGPT handles Thai queries well in GPT-4 and later; cites Thai-language sources when they exist
• Google AI Overviews is rolling out Thai-language coverage in 2026; coverage is partial as of writing
• Perplexity handles Thai queries but its source corpus is more English-weighted; Thai-language sources get cited less proportionally
• Claude with search handles Thai queries well but is used less than ChatGPT in Thailand

The practical implication for Thai businesses: Thai-language content gets cited less in AI search than English content because the AI engines' source corpora are still English-skewed. This will normalise over time.

Auditable changes for AI-search readiness:

1. Add author bylines to every blog post. If you don't have a real expert author, attribute to the company team page with credentials. 2. Add Article schema with dateModified on all content. Update the dateModified when you actually update content. 3. Add FAQ sections to every service page with 5-8 questions. Include FAQPage schema. 4. Find your three most important blog posts and rewrite the opening paragraph with concrete facts, numbers, and dates. AI engines extract from the opening; vague opening = no citation. 5. Check if your site appears in Google AI Overviews for queries you'd expect to rank for. If competitors appear and you don't, that's the citation gap to close.

AI search will be 30-50% of search traffic by 2028. Investing in AI-friendly content patterns now (schema, bylines, specificity, FAQ structure) is not a wild bet — it's just good SEO that compounds.

Ignoring it costs you visibility in the channels your customers are increasingly using. Over-investing in "AI SEO" agencies that promise special techniques is wasted money.

The boring middle path — solid SEO with the additions above — wins.