Thailand Lacks 80,000 AI Specialists — How SMEs Can Use AI in ERP

1. Thailand's AI Talent Gap — What's Happening

In May 2026, Thailand's Ministry of Higher Education, Science, Research and Innovation (MHESI), together with True Corporation, Google, and 20+ universities, launched "AI for All Thais." The ministry framed it bluntly: developing Thailand's AI workforce is not merely an education issue — it's a competitiveness issue. Without sufficient AI talent, major technology companies may choose to invest elsewhere in the region.

Number	Meaning	Implication for SMEs
~80,000 specialists	Estimated AI talent shortage today	Hiring an AI engineer is nearly impossible for SMEs
20 million	Target population for AI upskilling within 4 years (up from 12M)	Future SME hires will arrive AI-literate
< 20%	Share of Thais with meaningful AI access in 2025	The market is wide open — early adopters win
Two phases	"AI First Citizen" (Gemini Academy) + "AI for Future Workforce" (45-hour credit-bearing curriculum)	Pipeline of trained workforce SMEs can hire from

Separately, the State of AI Engineering 2026 report finds that Thai organizations adopt AI faster than most ASEAN peers — but lag on operational system readiness. In other words, executives are eager but back-end systems aren't ready. This is the gap the previous AI Adoption Gap article highlighted — and it's the gap that ERP systems with built-in AI can close.

2. Why SMEs Can't Win the AI Hiring War

Before discussing solutions, you need to understand why "just hire an AI engineer" doesn't work for SMEs. There are four reasons:

Factor	Big Tech / Banks	Typical SME
AI Engineer salary	2–3× regular dev salary + bonus + stock	Cannot match the package
Project appeal	LLM, RAG, large-scale fine-tuning	Mostly basic automation — top engineers get bored fast
Data infrastructure	Data lake + dedicated data team	Data scattered across Excel and legacy systems — no fuel for the engineer
Retention	Career path + AI lab + research	Engineer leaves quickly — typically back to Big Tech

The pattern we see all the time: an SME hires "an AI engineer," spends six months on a proof-of-concept, the engineer resigns just as it goes live, and the system becomes unmaintainable. 1–2 million baht of investment becomes sunk cost. This is exactly why the AI Investment ROI article argues that for SMEs, "buying AI features" beats "building AI" in nearly every scenario.

3. Five Ways SMEs Can Use AI — Without Hiring an AI Engineer

The answer for SMEs is neither "don't use AI" nor "hire an engineer." It's a shift in mindset — from 'build AI ourselves' to 'use AI the vendor built for us.' Five paths are available right now:

Path	What It Is	Who Uses It	Starting Cost
1. AI in ERP	AI features built into your ERP — auto-categorization, anomaly detection, AI assistant	Accounting, procurement, HR	Bundled in ERP fee
2. SaaS AI tools	Claude, ChatGPT, Gemini — for legal review, copywriting, data analysis	Any read/write-heavy team	$20–100 / seat / month
3. Vendor-provided AI agents	Domain-specific agents — e.g., Claude Finance Agents for finance work	CFO, analysts	License + setup
4. Train existing staff	Teach accountants/HR/sales to use AI as an "assistant" — not a replacement	Current employees	2–5K baht / person / course
5. Vendor partnership	Engage vendors that bring AI capability via the Enterprise AI Firm model	Executives (decision)	Project-based

The critical observation: none of the five paths require a full-time AI engineer. They all rely on "AI that already exists" and ask employees to "use it well" rather than "build it." The distinction matters enormously: building an LLM requires 5–10 senior data scientists plus a GPU cluster worth hundreds of millions of baht; using AI requires one employee with a Claude or ChatGPT login.

4. AI Features Your ERP Should Have — An Executive Checklist

When evaluating an ERP with built-in AI — or a vendor partnership — executives should ask about these capabilities before signing:

AI Feature in ERP	Use Case	Measurable ROI
AI financial summarizer	Summarize P&L, cash flow, flag anomalies for executives	50–70% reduction in monthly review time
Anomaly detection	Flag unusual transactions or stock movements — see AI in Accounting	Catch fraud/errors before close
Auto-categorize transactions	AI assigns GL codes from vendor descriptions	30–60% less data entry
AI assistant for queries	"What's this month's revenue?" "Which SKUs are running low?"	Executives self-serve insights
Predictive forecasting	Demand, cash flow, invoice due-date forecasts	20–40% reduction in over/under-stock
Document AI	Auto-ingest invoices, POs, tax receipts	70–90% less manual key-in

Notice — none of these six features add headcount. They live inside an ERP your business already runs: no new engineer to hire, no data team to stand up. The decisive choice is picking a vendor that ships these features. AI + ERP in 2026 is becoming the new industry baseline.

5. A 6-Month Plan — How an SME Starts Using AI Without Heavy Investment

Instead of betting on an AI Engineer hire and waiting 6–12 months for a POC, this plan delivers visible results every 30 days — with no new headcount:

Total six-month budget — for a 30–100 person SME, expect to land in the low-six-figure baht range (SaaS licenses + workshops). That's less than one month's salary for an AI engineer — and the impact spreads across the whole organization rather than sitting on one engineer's laptop who might leave next quarter.

6. Four AI Misconceptions SMEs Should Avoid

Across SME advisory work over the past year, four recurring misconceptions consistently lead to wasted investment:

Misconception	Why It's a Problem	Right Frame
1. We must hire an AI engineer first	Long hiring cycle + expensive + hard to retain	Start with SaaS + AI in ERP — hire later if needed
2. AI requires a big data team	Most SMEs don't have enough data even to train a model	Use vendor pre-trained models — no training required
3. AI = full automation	Wrong expectation — when AI errs, users abandon it	AI = augmentation (assistant), not 100% automation
4. AI doesn't need governance	Customer data leaks → PDPA exposure + reputational damage	Set an AI policy from day one — who can use what, with what data

Misconception #4 is the most consequential: even without an AI engineer, SMEs need an AI policy stating which AI tools employees may use, what data they may submit, and how outputs are stored. The moment a staff member pastes customer records into a free-tier ChatGPT, legal risk is created instantly.

7. The Vendor's Role — What Executives Should Ask Before Signing

Because SMEs depend on vendors as their "AI intermediary," vendor selection matters more than any hire. Use these questions when evaluating an ERP vendor:

Question	"Pass" Criteria
1. What AI features does your system already have — not on the roadmap?	Live today + demo-able right now
2. Whose LLM do you use? Where is data stored?	Clear data residency policy + no unnecessary cross-border transfer
3. Does your implementation team know AI workflows?	Real case studies + has shipped AI for clients
4. Can you train our staff — in Thai?	Curriculum + Thai-speaking trainer + Thai documentation
5. If the AI feature breaks, who owns it?	Clear SLA + Thai-language support

Question #5 matters enormously. A vendor that sells "AI in ERP" but, when something breaks, says "you'll need to hire a consultant" was selling AI marketing — not AI you can actually use.

Summary

Topic	Key takeaway
Thailand's AI specialist gap	~80,000 — Big Tech and banks absorb the supply; SMEs shouldn't try to compete
"AI for All Thais"	20 million in 4 years — future SME hires arrive AI-literate
Five SME AI paths	AI in ERP, SaaS AI, vendor agents, train staff, vendor partnership
Who uses AI in an SME	Accounting / HR / sales — not an AI engineer
Six-month plan	SaaS Month 1 → train Month 3 → ERP AI Month 4 → expand Month 6
Vendor questions	Real features today + data location + Thai-language training

"Thailand's 80,000-person AI shortage is not the SME's problem — it's a problem for Big Tech and banks competing for talent. SMEs that adopt 'vendor AI' through ERP, SaaS, and partnerships are not at a disadvantage — they have an advantage. They don't pay top-of-market salaries, don't fight retention battles, and don't bear the risk when an engineer walks out the door. The ERP vendor's job is to make AI easy to use. The SME's job is to pick the right vendor."

References

• Bangkok Biznews — AI for All Thais: Thailand's 80,000 AI Specialist Shortage (May 2026)
• Naewna — AI Economy column
• Daily News — AI Talk 2026