Know Your AI, Question It and Keep Your Nerve, Thai Investors told

 

Liberator's Watanya Bunnag urges investors to question AI and stay calm; Finnomena's Jessada Sookdhis sees the tech rout as a buying chance.

 

Investors should use artificial intelligence as a pilot, not an autopilot, a leading Thai securities executive said on Tuesday (20 October 2026), warning that over-reliance on AI tools can lead people to buy at market peaks.

 

Watanya Bunnag, Co-CEO of Liberator Securities, was speaking at Nation Forum 2027: Thailand's Agenda, in the "Fintech in AI Era" session of the "Money in the Air 2027" panel.

 

She shared the stage with Jessada Sookdhis, CEO and co-founder of Finnomena Group.

 

Both said the recent sell-off in AI stocks was more likely a buying opportunity than a bubble bursting, but both cautioned about the risks of misusing the technology and of AI-enabled scams.

 

 

Same tools, different edge

With everyone able to use AI, Watanya said an investor's edge lies in three things. First, investors must know which AI they are using. Most frontier tools are large language models built on probabilities, which can echo prevailing sentiment.

 

She recounted a client who said AI had led her to buy near a market peak. Such tools excel at gathering news, compiling analysis and reading financial statements, she said, but humans must stay in control.

 

Second, investors should ask specific, probing questions and have AI debate the risks rather than flatter them. As an example, they could ask how their portfolios would be affected if China stopped easing rare-earth restrictions on 10 November.

 

Third, access to markets is unequal. Thai equities alone may be too narrow to capture where AI money is flowing, including the United States, Taiwan and South Korea.

 

 

 

Investors must also keep their nerve, she said. Stocks that fell nearly 20 per cent last week partly rebounded on Monday, and those who sold in panic locked in losses. AI should be used to assess risk in advance, with humans making the final decision.

 

 

Why the market fell

Watanya noted that chip stocks in the United States and South Korea fell sharply last week, and that warnings of an AI bubble have circulated for a year. She said the situation differs from the dot-com era, when companies borrowed heavily and revenues failed to follow.

 

Earnings at firms such as Nvidia and TSMC show real demand, she said, and businesses are only beginning to embed AI agents to raise efficiency and cut costs. She remains positive but urged investors to favour strong companies rather than those with excessive debt.

 

She also cited remarks at the recent IMF and World Bank meetings that the global economy is being propped up by AI, and that without it world GDP would be about 1 percentage point lower.

 

Jessada said the causes were not a fall in AI use. The first leg down in July, which hit Samsung and SK Hynix, was driven by excessive speculation, with South Korean investors using heavy margin loans and facing forced selling.

 

This time, he said, AI infrastructure names fell further for different reasons. Newer "neo-cloud" firms, some of them former crypto miners, had raised capital to sell computing power measured in gigawatts, but higher interest rates tightened credit.

 

Nvidia's purchase guarantees apply only once facilities are built, and some projects stalled for lack of power, while the US–Iran conflict pushed oil and borrowing costs higher.

 

 

 

He said demand remains strong. AI has existed for about four years and agentic AI for around six months, he said, with AI systems themselves expected to become major users.

 

He compared the moment with the coal era, when a new energy source triggered a wave of change. When the audience was asked by show of hands whether the AI theme had burst or was a buying chance, almost the entire room chose the latter.

 

 

Three levels of use

Jessada described three levels of AI use. The first is AI as consultant, answering questions. The second is having AI build things. He said he uses an AI agent to read hundreds of research reports each morning and produce a briefing, and that such systems can be built through "vibe coding" in about 10 minutes. Few Thais have started, he said.

 

The third is agentic AI. He runs six or seven agents, including a bot that reads global news from 6 pm daily and summarises it, saving him one to two hours each morning.

 

He spends 40,000 to 50,000 baht a month on AI, which he said is worth more than hiring two or three staff. Skills, he said, can now be bought, so ideas will win.

 

He expects investors to split into those who adopt AI and those who do not, with the latter risking the fate of Kodak and Nokia. He coined the term "AI Alpha Investor" for those who use the tools well, noting that revenue per employee at S&P 500 companies is rising.

 

 

Scams outpacing regulators

Both warned of AI-enabled fraud. Watanya said Thai victims lose an estimated 30–40 billion baht a year, and that investment-scam complaints in the first quarter were four times the previous year's level.

 

Fraudsters can use AI to make deception faster and more convincing, and regulators struggle to keep up, she said.

 

The warning comes days after the IMF and World Bank Group, together with the Bank of Thailand, launched the Bangkok Blueprint on 14 October at their annual meetings in Bangkok.

 

The framework sets out 12 policy considerations to improve international cooperation against digital financial fraud.

 

Even so, Jessada said, Thailand remains vulnerable. He described a doctor who lost several million baht after a fake AI-generated trading board appeared in a Line investment group of apparently credible members.

 

 

Thailand's position

Asked how Thailand benefits, Watanya said capital flows first to chipmakers and then to AI model developers. Because those markets are expensive, money is spilling into places such as Thailand, where price-to-earnings ratios are low, and foreign funds have returned this year.

 

The Board of Investment has received record applications of around 1.4 trillion baht, about 1.1 trillion of them for data centres. These are applications rather than actual investment, she stressed, and Thai GDP grew only 1.9 per cent in the second quarter.

 

She said the benefits are not reaching Thais, and the country must choose between being a toll road earning fees for power, water and land, or an owner capturing returns from AI.

 

Jessada said new AI giants are only a few years old. He urged support for Thai start-ups, including solo founders using AI to do the work of teams, and hoped successful ventures would bring venture capital back to the country.