Thailand is changing quickly. AI is changing even faster. And law firms, including this one, have to adapt to both. That is the thread running through all three pieces this month: the biggest crackdown I have ever seen in this country, with billions of baht at stake; why AI has stopped being optional for lawyers; and a preview of the membership we are building for you.

1. The End of the Nominee Era? Thailand’s Biggest Crackdown in Decades

Last month I promised to write about protecting your home as a foreigner in Thailand. I am keeping that promise, but the news forced me to widen the frame. Because before we talk about the structures that work, we need to talk about the structure that is currently collapsing in real time: the nominee company.

I have lived in Thailand for about 22 years and worked in the legal industry for about 20 of them, and I have watched this country run enforcement campaigns before. The border run crackdowns, three or four of them since I arrived, each one ending the game for people living here on perpetual tourist stamps. The surrogacy crackdown of 2014 and 2015, which shut down an entire industry almost overnight once the scandals hit the international news. The student visa crackdown around 2024, when immigration finally went after the language schools selling ED visas to people who never attended a class.

Each of those was serious for the people caught in it. None of them compares to what is happening now. In my opinion, this is the biggest crackdown I have ever seen in Thailand, and the stakes are not measured in visa fees or overstay fines. They are measured in billions of baht of property. This is not a press conference campaign. This is arrests, disposal orders, and investigators pulling ten-year-old files out of accounting offices.

What a nominee is, in one paragraph

A nominee arrangement is when Thai citizens hold shares or land in their own name, but really on behalf of a foreigner, to get around the Land Code and the Foreign Business Act. The classic version: a foreigner wants a villa, a lawyer sets up a Thai company where Thai shareholders hold 51% on paper, the foreigner controls everything and pays for everything, and the company buys the land. Everyone signs, everyone smiles, and everyone pretends the Thai shareholders are real investors.

They never were. And here is the part I want to be blunt about: this was never legal. Not in a grey zone, not “tolerated,” not “technically fine if structured well.” It was illegal the day it was signed. What changed is not the law. What changed is that Thailand has decided to enforce it.

What I saw over 20 years

When I lived and worked in Nakhon Ratchasima years ago, nominee companies were rarely used. Korat is a big city but a small world; the local officials knew who was who, everyone could see what was happening, and nobody wanted to put their name on it. The nominee industry did not grow in places like that. It grew where the money was: Phuket, Koh Samui, Pattaya, and the other resort destinations, where thousands of foreigners bought villas through company structures and an entire ecosystem grew up to serve them.

And let me correct something you will read in the foreign press: this was never just a foreigner problem. Thai shareholders signed the share certificates. Thai accountants filed the annual accounts, year after year, for companies with no revenue and one villa. Thai lawyers drafted the structures. Thai developers marketed them openly in their sales offices. Thai citizens owned the shares, often owned the land, and profited from every step of the arrangement. If the era is ending, it is ending for all of them, not just for the foreigner in the villa.

Our firm made a deliberate decision, many times over the years, to walk away from this work. I will not pretend it was always easy; we watched competitors take fees we refused. But I always knew these structures were illegal, and I was not going to build a firm’s reputation on arrangements designed to fail.

You do not have to agree with the policy to understand it. Thailand restricts foreign land ownership for the same reason many countries protect farmland, coastlines, or strategic industries: it wants Thai land to remain in Thai hands. A nominee structure is not a clever workaround in the government’s eyes. It is a fraud on that policy, dressed in corporate paperwork. Once you see it from that side of the desk, the current campaign stops being surprising. The only surprise is that it took this long.

The phone call that should worry you

About a month ago, a former client contacted me. He had been arrested in Koh Samui. Here is what makes his case important: his company had been closed for several years. The property had already been sold. In his mind, and honestly in most lawyers’ minds until recently, the matter was finished.

It was not. Authorities investigated an accounting firm, found historical documents in its files, and began pursuing old nominee cases from the paper trail. His closed company was in that paper trail. Closing a company does not shred its history; the share registers, the transfers, the annual filings all still exist, sitting in government archives and accountants’ cabinets. Penalties vary enormously depending on the circumstances, such as cooperation, the age of the case, and whether the property is still held. But the exposure does not simply evaporate because you dissolved the company.

He is not alone. Other previous clients have been instructed to dispose of properties within a short period. Structures that sat unexamined for a decade are being reviewed aggressively. The environment today is simply not the environment of ten years ago, and anyone still making decisions based on how things worked in 2016 is making decisions based on a country that no longer exists.

What actually works instead

Here is the frustrating part, as someone who has spent 20 years in this industry: legal alternatives have existed the whole time.

• สิทธิเก็บกิน. A registered right to use and enjoy property, potentially for life. Registered on the title deed at the Land Office, enforceable, and completely legal for foreigners.
• พื้นผิว. The right to own the building separately from the land. You cannot own Thai land, but you can own the house standing on it.
• A properly registered 30-year lease. Thirty real years, not the 90-year fantasy some developers still sell, which the Supreme Court has already dismantled.
• Genuine Thai ownership. Often through a real marriage, which brings me to a distinction that matters.

A Thai spouse who genuinely owns property is not a nominee. If your wife owns the land, controls it, and it is truly hers, including if the marriage someday ends, that is Thai ownership, exactly what the law intends. A “spouse” arrangement built with side agreements, undated transfer documents, and the understanding that she owns nothing in reality? That is a nominee arrangement with a ring on it, and the courts can see the difference. The test is always substance: who paid, who controls, who truly benefits. See our guides on ข้อจำกัดการถือครองหุ้นโดยชาวต่างชาติ และ usufructs for married couples.

If you are forming a company in 2026: legitimate foreign-run businesses can still register Thai companies, and we do it regularly. But expect scrutiny that did not exist two years ago. Shareholding structures are examined, Thai shareholders must show real sources of funds, the company must have genuine business activity, and registrars are trained to spot nominee indicators. The 2026 FBA amendments also opened more sectors to full foreign ownership, which removes the excuse for nominees in many cases. Details in our guide to setting up a company in Thailand as a foreigner, and more on the enforcement wave in our articles on the 2026 nominee crackdown และ the law and court cases behind it.

2. Why AI Is No Longer Optional for Lawyers

When we restarted and reorganized this law firm around AI in 2023, a lot of people thought it was a mistake. Some said it politely. Some did not. Fellow lawyers did not understand what we were trying to build. Why would a small firm in Thailand tear up its workflow and rebuild everything around a technology that, at the time, mostly wrote bad poetry and invented facts with total confidence? Many were certain AI would never be reliable enough for serious legal work.

Three years later, I think it is fair to say those assumptions have not aged well.

The timeline, as I lived it

• 2023. Most people treated AI as a novelty. A toy for writing emails. Professionals dismissed it, and honestly, on the raw output of the day, I understood why. We bet on the trajectory, not the snapshot.
• 2024. People started realizing the technology was serious, and then spent the entire year focused on the problems. Hallucinations, confidentiality, the lawyer who filed fake cases. All real concerns. But “it has problems” quietly became an excuse to not learn it at all.
• 2025. The tools became genuinely useful. For some of us, they became indispensable. This is the year the gap opened between professionals who used AI daily and professionals who had tried it once in 2023 and formed a permanent opinion.
• 2026. The newest models, like Claude’s Opus and Fable-class frontier models or advanced Chinese models such as GLM 5.2, operate at a level that would have seemed impossible three years ago. But the models are only half the story. The real shift is agent systems and database-connected workflows. You no longer work with a lone chatbot; you architect a system and connect it to your emails, your calendar, your document templates, your legal database, your internal knowledge.

I will not bore you with the technical details, because the technical details are not the point. The point is what comes out the other end.

“But everything I see from AI is garbage”

I hear this constantly, and I understand where it comes from. The internet is drowning in what people call AI slop: generic, confident, wrong, and identical to a million other pages. If that was my only exposure to AI, I would dismiss it too.

But here is what I need you to understand: the problem is not AI. The problem is poor implementation.

Compare two setups. On one side, a basic chatbot with no database: it knows nothing about Thai law beyond whatever fragments landed in its training data, so it fills the gaps by guessing, and it guesses fluently. On the other side, a legal system connected to structured databases of actual Thai statutes and Supreme Court decisions, real legal documents, internal firm knowledge built over 20 years, and specialized workflows with a lawyer verifying the output. These two things share a technology and nothing else. The difference is not 20% better. It is a different profession.

Judge it yourself

This newsletter comes with two legal research reports, both anonymized. Both deal with the same subject: a foreigner’s villa investment held through a lease and sublease structure on an island in southern Thailand. One is a report the client received elsewhere. One is ours.

Read the other report first (download it here). It looks professional at first glance, and by its form I am fairly confident it was produced with AI. But it is general advice. It does not go deep. It does not explain the title deed. It refers to a long list of important documents without analyzing how any of them was actually drafted or presented. It is exactly the poorly implemented AI work I described above: the technology was available, the system behind it was not.

Then read ours (download it here). It works through the title deed and its registration history, the head lease, the registered sublease, the construction permit, and the buyback, and it tests the whole structure against specific Supreme Court decisions. It was produced with our research system, which analyzes a database of 140,000 Supreme Court decisions and laws of Thailand with tools we built for exactly this work. That system is now the engine of our Thai legal research service.

Same underlying technology. Two very different results. That is my whole argument, in two attachments. You can clearly see the difference in quality yourself.

Any good lawyer should use AI

After last month’s newsletter, a reader wrote to tell me I was mixing up AI and large language models. Not quite, but I welcome the debate: an LLM is one type of AI, the type driving the current wave, and I use the broader word on purpose because the systems I describe are more than a language model alone. What that email really tells me is that readers are now engaged enough to argue the terminology. Two years ago, nobody was.

Another reader told me I should stop talking about AI, because lawyers should focus on law.

My answer is the title of this section: any good lawyer should use AI. A lawyer refusing to learn it in 2026 is making the same mistake a doctor would make by refusing modern diagnostic tools. Imagine a doctor who says: “I don’t trust MRI machines. I palpate, I listen, my professor taught me this way in 1995.” You would not admire his purity. You would find another doctor. Not because he lacks knowledge, but because he refuses to use the best available tools to apply that knowledge to your problem.

Let me state my position clearly, because I have thought about it for three years and I am no longer hedging: AI will not replace every lawyer. But lawyers who use AI will replace many lawyers who refuse to use it. The future belongs to professionals who combine experience, judgment, and legal knowledge with AI systems. Remove any one of those and it falls apart. Experience without AI is slow and expensive. AI without judgment is dangerous. I have caught AI systems inventing Thai Supreme Court decisions with perfectly plausible case numbers, and only 20 years of experience let me smell that something was wrong. The combination is where the value is.

How we actually use it

Inside ThaiLawOnline, AI now touches almost everything: research across our legal databases; first drafts of documents that a lawyer then reviews and takes responsibility for; knowledge management, so that what we learned solving a case in 2019 is findable in 2026; quality control, cross-checking documents for inconsistencies; translation between Thai and English, verified by bilingual staff; and workflow automation for the repetitive administrative work that used to eat entire afternoons.

Notice what is not on that list: AI does not decide anything. It does not sign opinions, advise clients unsupervised, or file anything. It assists professionals; it does not replace them. Every output that matters passes through human judgment. That is not a limitation of our system. It is the design.

3. The Future of ThaiLawOnline Membership

Last month I told you we were bringing back membership and promised more details. This is that article, and we have now published a full explanation on the website: explaining our membership. Let me be clear about what this is not: it is not a sales pitch. There is nothing to buy today. This is a preview of where the firm is going, because you are the people it is being built for, and I would rather show you the workshop than the showroom.

An old idea whose time finally came

The membership idea has existed for years. I have sketched versions of it on and off for most of a decade: a way for people living in Thailand to get reliable legal documents and information at a fair price, without paying full consultation rates for things that should not require them.

Every time, I shelved it. Not because the idea was wrong, but because building it properly was impossible for a firm our size. Document generators that handle real legal logic, a knowledge base that stays current, tools that work in two languages: that was enterprise software territory, with enterprise budgets. What changed is the same thing I wrote about in the previous article. The AI systems we built for our own practice turned out to be exactly the infrastructure a membership needs. We are not building a membership from scratch; we are opening a door to the workshop we already built.

Why most legal memberships are disappointing

I want to be honest about the industry, because you have probably seen these offers before. Most legal membership products share the same problems: high fees for what is delivered; limited value, usually a folder of generic templates you could find free elsewhere; unrealistic promises; and my personal favorite, the “unlimited consultations” that turn out to be very limited indeed once you read the fine print or try to book a second call.

The model is usually marketing-first: sign people up, deliver as little as possible, count on them forgetting to cancel. I have no interest in running that business. If I cannot build something I would honestly recommend to a friend living in Thailand, I will not launch it.

What we are building instead

• Document generators. Not blank templates, but guided tools that produce real documents: last wills, loan agreements, powers of attorney, usufruct agreements, employment contracts, and other documents people in Thailand actually need.
• Contract and legal templates with plain-language explanations of what each clause does and when you need it.
• Business resources for company owners navigating exactly the kind of compliance environment I described in the first article.
• AI legal assistants connected to our legal databases. Not a generic chatbot, but systems that answer from actual Thai law. Specialized versions for businesses and for individuals, because a restaurant owner in Phuket and a retiree in Chiang Mai do not have the same questions.

Substantial work is already done. New template systems are built. New databases are running. New AI tools exist and are being tested internally. New member sections of the website are constructed. I am telling you this so you know it is real, and I am equally telling you that development is ongoing and not fully completed. I will not launch it half-baked to hit a marketing date. When it opens, it will be because it works.

The research layer underneath

Let me show you what makes this different, using one example: usufruct. Most websites, including, frankly, most law firm websites, offer a generic article: what a usufruct is, that it is registered at the Land Office, that foreigners can hold one. What we have been building goes much deeper: annotated Supreme Court decisions on usufruct, drawn from large collections of judgments in our database, showing how courts actually rule when a usufruct is challenged by heirs, in divorces, or after a sale of the land. Research of this kind is simply not commonly available online, in Thai or English.

That is the difference between structured legal knowledge and one more generic article. When a member asks our AI assistant about protecting their home, it will not answer from a blog post. It will answer from the case law.

Now also in Thai

One more development I am genuinely proud of: the website is now available in Thai. It might seem strange that a firm serving foreigners spent months on a Thai version. It is not strange at all. It reaches a far larger audience. It improves legal education in a country where accessible legal information in plain Thai is rarer than you would think. And most importantly: nearly every foreigner we work with has Thai people in their life. A spouse, in-laws, business partners, staff. When a foreign husband reads our usufruct guide in English and his Thai wife reads the same guide in Thai, they can make a decision together, from the same information. For mixed Thai-foreign families, that matters more than any feature we could build.

I will finish with something personal. The last few months have produced some of the most exciting developments in this firm’s history, and I say that having been at this for 20 years. Not because of marketing. Not because of growth numbers. Because the quality of the legal information and the legal tools has improved dramatically. Whatever the membership becomes commercially, the thing itself, the actual legal substance, is the best work we have ever done.