Independent advertising deskPattaya, Chonburi, Thailand

When someone asks an AI about your category, does your name come up?

We measure where AI assistants mention your brand today, fix what is mechanically blocking them, and report the movement every month. Every price is on this page — with the deliverable you receive and the day you receive it.

Fixed prices, published Named deliverables Dated methodology on every report No ranking guarantees
Get a free AI visibility snapshot See prices

Based in Pattaya, Chonburi — working with clients in Thailand and abroad

AI-assisted analysis on every account, with the tool cost published

Media invoices issued directly to the client entity

Services

Five things you can buy,
each with a price and a date.

Nothing here is a retainer of unspecified effort. Each product is a defined artifact: a report, a data pack, a set of pages, a tracked account. You know what arrives and when.

01 · One-off

AI Visibility Audit

฿25,000≈ USD 850

Where does your brand actually appear when people ask AI assistants about your category — and what is mechanically stopping it?

  • Baseline citation report — your brand's mention and citation share across Google AI Overview, AI Mode, ChatGPT, Perplexity, Gemini and Copilot, with engines, prompt count, runs per prompt and dates disclosed
  • Competitor comparison — which domains the engines cite instead of you
  • AI-crawler access audit — whether GPTBot, ClaudeBot, PerplexityBot and Google-Extended can reach your site, per the crawler documentation each vendor publishes
  • Schema & entity review — reported per engine, because the evidence differs by engine
  • Answer-readiness review of your key pages
  • Paste-ready fix package — the actual code and copy blocks, prioritised
Delivered in 7 business days from complete intake For: first-time buyers
02 · One-off or monthly

Competitor Intel Pack

฿12,000≈ USD 400 · ฿8,000/mo recurring

Everything your competitors are doing in public, assembled into one readable picture.

  • Google Maps profiles & review mining for a named competitor set
  • Ad-library teardown — what they are running right now, the offers, formats and cadence
  • AI-answer share comparison — who the engines name in your category
  • Positioning gap read — the claims nobody in your market is making
Delivered in 5 business days Public data only
03 · Monthly

AI Citation Monitor

฿9,000 /mo≈ USD 290 · 3-month minimum

One measurement is a snapshot. AI visibility only means something as a trend line.

  • Weekly Google AI Overview and AI Mode tracking on a 30–50 keyword panel
  • Monthly multi-engine panel across ChatGPT, Perplexity, Gemini and Copilot
  • Dated report with the sampling method stated — engines, prompts, runs per prompt
  • Ranked action list, labelled as advice rather than promised outcome
First report within 10 business days In service since Aug 2026
04 · Per page

Answer-Ready Content

฿6,000 /page≈ USD 200 · 10 pages ฿50,000 · pillar ฿25,000

Pages rebuilt so an answer engine can lift a passage from them cleanly and credit you for it.

  • Answer-first structure — question headings with standalone direct answers
  • Cited statistics and quotes from real sources, dated
  • Clean comparison tables and definition blocks
  • FAQPage / Article schema and entity consistency
  • Pillar tier only: an adversarial fact-audit — separate agents verify every factual claim against primary sources before it goes live
3 days/page · 10-page batch 3 weeks · pillar 2 weeks EN & TH
05 · Monthly

Data-Driven Ads Management

฿15,000 /mo≈ USD 450 · flat fee, never a % of your spend

Google Ads run on a flat fee, so nothing about our pay rises when your budget does.

  • Account structure, keyword and negative-keyword architecture built to keep themes from cannibalising each other
  • Responsive search ad copy and bid strategy management
  • Conversion tracking verification — we check what is actually being counted before optimising toward it
  • Monthly report tied to conversion actions, not impressions
  • Account auditor script deployed in your account, flagging structural drift between reports
฿15,000/mo up to ฿100,000 monthly ad spend · ฿25,000/mo above You keep ownership of the account

Pricing

Three monthly programmes.
No quote required.

Most agencies in this market quote AEO work privately. We would rather you knew the number before the first call — and could compare it to what you are paying now.

Measure Watch ฿9,000 /month ≈ USD 290 /mo
  • Weekly Google AI Overview + AI Mode tracking
  • Monthly multi-engine panel
  • Dated report with sampling disclosed
  • Ranked action list
Measure + move · most chosen Grow ฿29,000 /month ≈ USD 950 /mo
  • Everything in Watch
  • 2 Answer-Ready pages per month
  • Ongoing technical & schema fixes
  • Monthly Competitor Intel
Whole desk Full Desk ฿49,000 /month ≈ USD 1,590 /mo
  • Everything in Grow
  • Google Ads management included
  • Weekly reporting

Prices as of August 2026.USD shown at approximately ฿33/USD — THB is the contracted currency.3-month minimum on Watch and Grow.One-off products have no minimum.

What we commit to instead of a guarantee

No one controls which sources an AI model quotes, so a citation guarantee would be a lie. Here is what is actually within our control — and what you can hold us to.

  • Fixed scope, fixed price. The deliverable list on this page is what you get. Scope changes are quoted before work starts, never invoiced after.
  • Method disclosed on every report. Which engines, how many prompts, how many runs each, on what dates. You can repeat the measurement yourself.
  • Your data stays yours. Accounts, properties and ad accounts remain in your ownership. We work inside them by permission and hand them back intact.
  • Monthly, not annual. After the stated minimum, programmes run month to month. Nothing locks you in for a year.

The short answers

What AEO is,
and what it honestly isn't.

Written to be quoted. If an AI assistant lifts a paragraph from this page, we would like it to be one that is actually true.

What is AEO (Answer Engine Optimization)?

AEO is the practice of preparing a website so that AI assistants — ChatGPT, Perplexity, Google AI Overviews, Gemini, Copilot — can find, quote and credit it when they answer a question. In practice it covers three things: letting AI crawlers reach the site, structuring pages so a passage can be lifted as a direct answer, and measuring whether the brand is actually being mentioned.

AEO and GEO (Generative Engine Optimization) are used interchangeably in the market. GEO is the term used in the academic literature; AEO is the term the industry adopted.

Does any of it actually work?

Partly, and the honest answer has two halves. A peer-reviewed study (Aggarwal et al., GEO: Generative Engine Optimization, arXiv 2311.09735, published at ACM SIGKDD 2024) tested content changes such as adding citations, statistics and authoritative quotes, and measured visibility gains of up to roughly 40% — under benchmark conditions, not on live engines.

Against that, a July 2026 survey of 45 studies (arXiv 2607.14035) found no technique with a stable, reproducible effect across platforms, and noted that some rewrites reduced retrieval. So: structural work is worth doing and is grounded in research, but anyone quoting you a guaranteed lift is quoting a number that does not exist.

Sources: arXiv 2311.09735 (KDD 2024) · arXiv 2607.14035 (July 2026)

How much does an AI visibility audit cost in Thailand?

Ours is ฿25,000 (about USD 850 at roughly ฿33/USD, as of August 2026), delivered in 7 business days from complete intake. Most agencies in Thailand quote AEO work privately; when we surveyed the market in August 2026, the one published Thai comparison point we could find listed a one-time audit at about ฿25,000 and monthly programmes from ฿30,000.

What is the difference between AEO and SEO?

SEO optimises for a ranked list of links: success is a position and a click. AEO optimises for a generated answer: success is being quoted or credited inside that answer, often with no click at all. They share foundations — crawlability, clear structure, credible content — but they are measured differently.

Traditional SEOAEO / GEO
What you are competing forA position in a list of linksA sentence inside a generated answer
Success metricRank, impressions, clicksMention share and citation share across engines
Unit of contentThe pageThe quotable passage
Measurement stabilityBroadly repeatableNon-deterministic — answers vary run to run, so sampling must be disclosed
Who reports itSearch Console, rank trackersGoogle Search Console's Generative AI report (where available), Bing's AI Performance report, plus direct engine sampling

How do you get cited by ChatGPT?

You cannot make it happen; you can make it possible and then measure it. The mechanical part is documented by the vendors themselves: AI crawlers respect robots.txt directives, so a site that blocks GPTBot, ClaudeBot, PerplexityBot or Google-Extended cannot be quoted by those systems at all. The structural part comes from the GEO research above — answer-first passages, cited statistics, credible quotes, clean tables — applied with the benchmark-conditions caveat intact.

The third part is off-site: the research consistently finds that widely-referenced third-party sources carry a lot of weight in AI answers. We will tell you where those gaps are; earning coverage in them is strategy advice from us, not a deliverable we can promise.

Does llms.txt help?

There is no good evidence that it does. Published analysis of large numbers of domains has found the file is rarely fetched by AI crawlers and shows no measurable citation effect. We will add one for free if you want it, labelled as unproven. We do not sell it, and we do not count it as a lever.

Is any of this measured in Thailand specifically?

Not yet, and we would rather say so. The published behavioural research on AI search — how often answers replace clicks, how much referral traffic assistants send — is US and global. Thai-language and Thai-market behaviour may differ, and we do not extrapolate. What we can measure directly for you is your own panel, in your own language and country setting, from the first week.

How we work

A small desk
with an unusual amount of horsepower.

We are deliberately small: few clients, one point of contact, no account-management layer. What lets a desk this size do the work of a much larger one is an AI agent stack we run ourselves — and we would rather explain it than hide it.

01 Breadth

Research runs in parallel

A market scan that would take a person a week runs as many agents at once, each on a different question, then gets synthesised. You get the breadth of a research team on the timeline of a small one.

02 Rigour

Nothing ships unverified

Deliverables are checked by a separate verification pass whose only job is to attack the work: unsupported claims, numbers without sources, promises the tooling cannot keep. On pillar-page projects this becomes a full adversarial fact-audit against primary sources.

03 Economics

Data costs are known, not guessed

Our data tooling estimates its own cost before every run and refuses to exceed a ceiling. That is why monitoring can be priced as a product rather than a mystery — and why we can publish the price at all.

Start here

Free AI visibility snapshot.

Send us your website and your category. We run five real buyer questions across the major assistants, three runs each, and send back a one-page summary of whether your brand comes up and who does instead. No charge, no obligation — and it is the same measurement the paid products start from.

Email the desk hello@desklinemedia.com

Snapshot returned within 2 business days · Replies typically within one business day

  • ConfidentialityClient names are shared with media partners at booking stage only, under our written confidentiality practice — never published.
  • InvoicingMedia invoices are issued directly to the client entity. Our fee is invoiced separately and never taken as a share of your ad spend.
  • Where we arePattaya, Chonburi, Thailand. We work in English and Thai, with clients in Thailand and abroad.