By Thanaphat Thinakul · CEO, IMAI Group
An AI visibility audit is a review of how often and how accurately AI assistants like ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews name your brand, which pages they cite, and whether AI crawlers can reach your content. You can run a rough version yourself in about 20 minutes, or pay for a deeper one. Here's how to tell which you need.
- A self-audit tests 15 to 25 real customer questions across four AI platforms (Source: Shadow, 2026).
- Audits score five signals: entity recognition, structured data, trusted citations, topical authority, and documented outcomes (Source: Globe and Mail / ACCESS Newswire, 2026).
- Citation rates swing from 100% to 20% on the same question between runs, so repeated rounds beat a single snapshot [ImVisible platform data].
- A paid one-time audit in Thailand costs around 25,000 THB (Source: Relevant Audience, 2026).
- Schema markup and llms.txt are not confirmed AI ranking factors per Google.
What Is an AI Visibility Audit and Why It Matters in 2026
An AI visibility audit measures how often ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews name a company, which pages they cite as sources, and whether AI crawlers can reach the site's content (Source: LoudFace, 2026). It answers one blunt question: when a customer asks AI for a recommendation, does your name come up?
Agencies sell this as an AI SEO agency or GEO service. The core idea is simple enough to check before you pay anyone. Updated for 2026.
Why AI answers matter alongside Google rankings
Ranking first on Google no longer guarantees the click. More buyers ask ChatGPT or read Google AI Overviews before they visit any site. If AI names three competitors and skips you, that traffic never reaches your page, even with your blue link at position one.
Who should run this check
Founders and marketing managers at Southeast Asian SMEs should run this first. If you suspect AI assistants recommend rivals instead of you, a self-check shows how wide the gap is before you commit budget to a formal audit.
GEO vs SEO vs AEO: What's the Difference
GEO vs SEO comes down to the target system. SEO optimizes for Google's ranking algorithm. GEO and AEO optimize for generative AI assistants and answer engines that summarize or cite content directly instead of linking to a ranked list. All three overlap, but each aims at a different surface.
| Approach | Target system | Goal |
|---|---|---|
| SEO | Google search ranking | Rank high in blue-link results |
| AEO | Answer engines and snippets | Be extracted as the direct answer |
| GEO | Generative AI (ChatGPT, Gemini) | Be named inside AI-written replies |
SEO: ranking in search results
SEO earns position in Google's organic list through content, links, and technical health. It still drives traffic. But it stops at the ranked list and leaves AI answers untouched.
AEO: being the direct answer
AEO is answer engine optimization, the practice of structuring content so AI assistants can extract it as a direct, citable answer. Clear definitions, tables, and self-contained paragraphs make a page easy to quote.
GEO: being named inside AI-generated text
GEO is generative engine optimization, the practice of optimizing content and entity signals so generative AI systems are more likely to name and cite a brand. Want the full picture? Compare SEO vs GEO strategies to see where each fits your plan.
How to Run Your Own AI Visibility Audit in 20 Minutes
A self-run AI visibility audit means picking 15 to 25 realistic customer questions and asking each one in ChatGPT, Perplexity, Google AI Mode, and Claude, then recording whether your brand is named, which competitors appear instead, and which pages get cited (Source: Shadow, 2026).
- Build your 15-25 query list. Write questions the way customers type them: "best [service] in Bangkok", "[category] recommendations", "is [competitor] any good". Mix branded and unbranded prompts.
- Test across four AI platforms. Paste each question into ChatGPT, Perplexity, Google AI Mode, and Claude. Use a fresh chat every time so memory doesn't skew the result.
- Score and log the results. For each answer, note three things: did AI name you, which rivals showed up, and which URLs it cited. Tally your share of voice as a simple percentage.
Quick tip: Run each prompt twice. Your name often appears once and vanishes the next time, which already proves the fluctuation problem. A free AI-visibility check on imvisible.tech speeds this up.
The Five Signals That Decide Whether AI Names Your Brand
AI visibility depends on five signal categories: entity recognition, structured data, trusted source citations, topical authority, and documented outcomes, according to a widely used audit framework (Source: Globe and Mail / ACCESS Newswire, 2026). Strengthen all five and your odds of being cited rise across every assistant.
- Entity recognition: AI understands what your brand is and does.
- Structured data: schema markup helps machines parse your pages.
- Trusted source citations: mentions on reputable third-party sites.
- Topical authority: depth of coverage on your subject.
- Documented outcomes: case studies, reviews, and proof.
Entity recognition and trusted citations
Entity recognition and third-party citations carry the most weight. When respected sites describe your brand consistently, AI learns to tie your name to your category and repeats it in answers.
Why schema and llms.txt alone won't fix visibility
Schema markup and llms.txt are not confirmed AI ranking factors per Google. Treat them as technical hygiene, not a magic switch. Many agencies oversell them. They help machines read your content, yet they never force AI to cite you.
Self-Check vs Paid AI Visibility Audit: Price and Scope Compared
A paid AI visibility audit in Thailand runs around 25,000 THB as a one-time fee and can include a 50+ prompt snapshot, competitor comparison, and a 30/60/90-day plan (Source: Relevant Audience, 2026; SEO Bangkok, 2026). A free self-check costs nothing but takes manual effort and lacks repeated-round tracking.
| Feature | Free self-check | Paid audit (~25,000 THB) |
|---|---|---|
| Cost | Free | ~25,000 THB one-time |
| Prompts tested | 15–25 manual | 50+ prompt snapshot |
| Repeated rounds | Manual only | Tracked over time |
| Action plan | None | 30/60/90-day plan |
What a free self-check gives you
A free self-check gives you a fast reality check: whether AI names you at all, and who it names instead. That is usually enough to decide if the gap is worth paying to close.
What a paid audit adds
A paid audit from a GEO agency adds scale and tracking. It runs more prompts, compares competitors, and hands you a prioritized roadmap (Source: SEO Bangkok, 2026; Relevant Audience, 2026). See how ImVisible tracks weekly AI-citation sampling for ongoing monitoring.
Common Mistakes When Auditing AI Visibility
The most common mistake is treating a single AI query as proof. Citation rates can swing from 100% to 20% for the same question between separate runs, which makes repeated-round testing essential rather than optional [ImVisible platform data].
Trusting a one-time snapshot
A single pass can flatter or scare you for no real reason. One run shows your name; the next drops it entirely. Test in repeated rounds and report an average, not a lucky hit.
Ignoring third-party source dependency
In ImVisible's Thai portfolio, most brand mentions in AI answers come from third-party sites, not the brand's own pages [ImVisible platform data]. Polish only your own website and you miss where AI actually sources its recommendations.
Assuming schema markup guarantees citation
A frequent error is adding schema and expecting instant AI mentions. Schema helps parsing but guarantees nothing, since Google has not confirmed it as an AI ranking factor. Earn trusted citations first.
You can also see weekly rank checks with DataForSEO alongside your AI tracking, or learn more about AEO and GEO services when you're ready to act.
Related: AEO services for Google AI Overviews & ChatGPT
Frequently Asked Questions
What does an AI visibility audit actually check?
An AI visibility audit checks whether ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews mention a brand, which sources they cite, and whether AI crawlers can access the site's content (Source: LoudFace, 2026). It shows your share of voice against competitors in AI answers.
How many queries should I test in an AI visibility audit?
A recommended query set covers 15 to 25 target questions run through ChatGPT, Perplexity, Google AI Mode, and Claude (Source: Shadow, 2026). Mix branded and unbranded prompts, and test each one more than once to catch run-to-run variation in the results.
Is there a GEO agency in Singapore?
ImVisible is a Bangkok-based SEO, AEO, and GEO service delivered remotely in English for buyers in Singapore and Southeast Asia. There is no Singapore office or entity, yet the service covers Singapore clients fully through remote delivery and reporting.
How much does a paid AI visibility audit cost?
A one-time AI visibility audit in Thailand costs around 25,000 THB, while other Thai providers price their AI search optimization audits per project rather than publishing a fixed fee (Source: Relevant Audience, 2026). Scope and prompt volume drive the final price.
Do schema markup and llms.txt improve AI visibility?
Schema markup and llms.txt are not confirmed AI ranking factors according to Google, so treat them as supporting technical hygiene rather than guaranteed visibility boosters. They help machines read your pages but do not force any AI assistant to cite you.
Why does the same AI question give different answers each time?
AI answers are inherently variable. The same question tested in separate rounds can show brand citation rates swinging from 100% down to 20%, which is why audits should use repeated rounds, not single checks [ImVisible platform data]. Average several runs for a reliable picture.
Disclaimer: Pricing and platform behavior described here reflect 2026 sources and may change. AI ranking factors are not officially confirmed by Google; treat all optimization advice as guidance, not a guarantee.