Ask a Thai business owner today whether they "use AI," and the answer is increasingly yes—a chatbot widget on the website, an AI writing tool for social captions, maybe a customer-service assistant plugged into LINE. By the numbers, that owner is part of a genuine boom. By a different, more specific number, they're also part of a group that's stuck.
AWS's "Unlocking Thailand's AI Potential 2026" report, covered by Forbes Thailand, found AI adoption among Thai organizations grew 34% year over year, reaching roughly 220,000 companies—43% of Thai businesses, up from 32% the year before. Among adopters, 71% reported revenue increases and 84% reported improved efficiency. That's a real, fast-moving adoption curve by any measure. But the same report splits those adopters by depth of use, and the split is stark: 74% operate at a "basic" level—chatbots, off-the-shelf tools, minor efficiency gains—while only 9% have reached advanced implementation: custom-built systems, multi-model approaches, or autonomous agentic AI.
It sounds like Thailand crossed the AI adoption line and the rest is just a matter of time. Actually, adoption and capability turned out to be two different races entirely, and the businesses winning the first one aren't automatically winning the second. This piece looks at what separates "basic" from "advanced" AI use in practice, why that gap tracks suspiciously closely with a data-readiness problem Thai SMEs already have on record, and what it takes to move from one side of the 74%/9% split to the other.
Adopting AI and being ready for AI are two different scoreboards. Thailand is winning decisively on the first one and still building the second.
The Number Behind the Number
Thirty-four percent growth to 220,000 companies is the headline. But AWS's own survey names the barriers behind why most of those companies stall at "basic": 56% cite a shortage of skilled talent, 48% cite budget constraints, and 41% cite unclear AI and digital regulations. Only 19% of Thai organizations describe themselves as having high or complete readiness for advanced AI adoption, and 51% now name AI skill gaps as a major implementation barrier—up from 47% the year before, meaning the gap is widening even as adoption grows.
None of those named barriers is "we don't have access to AI tools." Access clearly isn't the constraint—220,000 companies already have it. The constraint sits somewhere else, in what a business can actually do once the tool is switched on.
The Advanced AI Ceiling: What Keeps Businesses Stuck at Basic
Call it the Advanced AI Ceiling: the point where a chatbot or writing assistant hits the limit of what generic, off-the-shelf AI can do without structured business data behind it. A "basic" AI tool can answer a general question, draft a caption, or handle a scripted FAQ—useful, and genuinely worth the 71% revenue lift many businesses report. An "advanced" AI use case—personalized recommendations, predictive follow-up, an agent that can actually act on a specific customer's history—needs something a generic model doesn't come with: organized, accurate, structured data about that particular business and its particular customers.
AWS's survey doesn't ask directly about data foundations, so this connection is our own reading of two separately verified data sets, not a claim either report makes on its own. But read alongside ETDA's finding that data analytics is Thai SMEs' single weakest digital-maturity dimension, the two numbers point at the same wall from two different angles: 74% basic AI use and a national data-analytics score dragging the whole digital maturity index down are consistent with each other, not coincidental. An AI tool can only be as personalized, as predictive, or as autonomous as the data a business feeds it—and a business that hasn't organized its customer data yet has nothing structured to feed an advanced use case, no matter how capable the underlying model is.
This Is the Same Gap, Different Report
That's worth stating plainly, because it changes where a business should actually spend its next budget. A business stuck at "basic" doesn't necessarily need a more advanced AI vendor. It may need the unglamorous work of getting its customer and sales data into one structured, reliable place first—the exact foundation an advanced AI use case would need to draw on regardless of which AI platform eventually sits on top of it.
Picture a hypothetical three-showroom furniture retailer in Bangkok called Somboon Furniture Gallery. Somboon installed an AI chatbot last year to handle delivery-time and care-instruction questions on its website—genuinely useful, genuinely counted among the 220,000 companies now "using AI" in Thailand. But the chatbot can't recommend a matching side table to a customer who bought a sofa six months ago, because nothing connects that chatbot to a structured record of what that customer purchased. Somboon isn't failing at AI. Somboon is missing the data layer that would let AI do anything more advanced than answer the same five questions on repeat.
That gap matters more than it might seem, because Thai consumers are already asking for exactly the kind of personalization Somboon can't deliver yet—see what Thai consumers actually want from AI, and why most SMEs can't deliver it for the demand side of this same ceiling.
The Data Foundations Playbook
- Audit What "Using AI" Actually Means for You—be honest about whether your current AI tool is answering generic questions or acting on specific customer data; most businesses in the 74% are doing the former.
- Structure Before You Automate—get customer and sales data into one consistent record before evaluating a more advanced AI feature; a smarter model fed disorganized data just produces confident-sounding mistakes faster.
- Advanced AI Needs Owned Data, Not Just Access—the ceiling most Thai businesses hit isn't which AI vendor they picked, it's whether they have structured data that vendor can actually work with.
- Budget for the Data Layer, Not Just the Model—of the barriers AWS's own survey names, talent (56%) and budget (48%) are usually spent entirely on the AI tool itself; some of that budget belongs in the data foundation underneath it.
A boutique law firm using an AI drafting assistant faces the identical ceiling—the tool can produce a generic contract clause instantly, but it can't flag that this particular client always negotiates a specific term differently, because that pattern lives in a partner's memory instead of a structured client record the AI could actually read.
Foundation: what this all comes down to
If there's one word for what actually separates the 9% using advanced AI from the 74% stuck at basic, it's Foundation—not which AI model a business chose, just whether the data underneath it is organized enough to make the model genuinely smarter instead of just faster at the same generic answers.
So here's the question worth sitting with: the AI tool your business already uses—could it recommend, predict, or personalize something specific to one of your actual customers today, or would it need a name and a fresh set of questions to even get started? If it's the second one, that's the Advanced AI Ceiling showing up in your own business, not a hypothetical one.
FAQ
Does 34% AI adoption growth mean most Thai businesses are AI-ready?
Adoption and readiness are different measures. Adoption grew to roughly 220,000 companies, but only 19% describe themselves as having high or complete readiness for advanced AI, and just 9% have actually reached advanced implementation.
What's the practical difference between "basic" and "advanced" AI use?
Basic use—chatbots, off-the-shelf tools, minor efficiency gains—describes 74% of adopters. Advanced use—custom-built systems, multi-model approaches, autonomous agentic AI—describes only 9%, and generally requires structured business and customer data the basic tools don't need.
Is the barrier to advanced AI mostly cost?
Cost is one factor—48% of Thai organizations cite budget constraints—but talent shortage (56%) and unclear regulation (41%) rank alongside it. Data structure isn't named directly in AWS's survey, but it's a prerequisite most "basic" AI tools don't require and most "advanced" ones do.
Should a business stuck at "basic" AI switch vendors to get more advanced results?
Not necessarily first. If the underlying customer and sales data isn't structured and reliable, a more advanced AI platform will struggle for the same reason the current one does—organizing that data is usually the higher-leverage next step.
What this looks like with NAWA CRM
NAWA CRM gives a business the structured layer an advanced AI use case would actually need: one customer timeline with conversations, decisions, and history in a consistent record, plus reporting that shows movement and conversion without rebuilding a spreadsheet by hand. It won't turn a basic chatbot into an autonomous agent by itself, but it's the foundation that decision has to sit on either way.
If your business already uses AI but isn't sure the data behind it could support anything more advanced, talk to us about what your current customer data setup can actually support.
Sources: Forbes Thailand, "องค์กรไทยใช้ AI พุ่ง 34% เพิ่มขึ้นเป็น 2.2 แสนบริษัท กว่า 71% รายได้เพิ่มขึ้น แต่ AWS เผยโจทย์ใหม่ มีเพียง 9% ที่ใช้ AI ขั้นสูง", citing AWS's "Unlocking Thailand's AI Potential 2026"