Betting big on small AI

Thailand is keen to leverage small artificial intelligence (AI) to create a big impact in many sectors.

Small AI solutions are focused applications designed to solve specific, practical problems, often using less data, computing power and money than large, general purpose systems. They could deliver significant gains in productivity and competitiveness, help narrow social inequalities and support economic growth.

Earlier Finance Minister Ekniti Nitithanprapas told reporters Thailand will study how to harness small AI to enhance the production and living quality of its farmers.

Nantapong Chiralerspong, director-general of the Trade Policy and Strategy Office, said the small AI concept is crucial for Thailand, especially for small and medium-sized enterprises (SMEs) that cannot afford to invest in big AI systems.

The Commerce Ministry wants to encourage SMEs in agriculture, tourism and manufacturing to use small AI tools to improve their operations, he said.

SMEs can use small AI to analyse market data, predict product demand, manage orders and communicate with international customers.

Thai policymakers were inspired to deploy small AI after the IMF-World Bank Group Annual Meetings in Bangkok last month where the technology was a widely discussed topic, covering the opportunities and challenges posed by AI, including concerns that automation could replace human workers and increase unemployment.

The World Bank backs small AI to bridge development gaps. With 2.2 billion people still offline and low-income countries accounting for less than 1% of ChatGPT usage, AI risks widening global inequalities.

The global lender supports small AI solutions as affordable tools that run on everyday devices to tackle local development challenges where infrastructure and digital skills are limited.

World Bank president Ajay Banga said the Thai tourism sector offers a model for generating employment and supporting broader economic growth through AI adoption, particularly by leveraging small AI solutions.

QUICK-WIN PROJECTS

Chanwit Boonchuay, president of the AI Entrepreneur Association of Thailand (AIEAT), told the Bangkok Post enterprises should start with business problems that deliver measurable results within a few months.

Small AI means affordable tools built for specific tasks and local conditions, including visual inspection, demand forecasting, document processing and compact language models. Some can run on existing phones or on-site devices.

"This suits Thailand, which must lift productivity as its population ages and firms face cost and competitive pressure. A factory can cut defects, a hotel can handle routine guest inquiries better, and a food business can reduce waste," Mr Chanwit said.

Medical AI should also be a priority. Hospitals and clinics can begin with tools for appointments, patient flow and documentation reviewed by professionals. The aim is to widen access to care while building Thailand's medical technology industry.

"For AIEAT, the immediate opportunity is in helping businesses adopt reliable tools adapted to Thai language and industry needs," he said.

Firms should choose the simplest AI system that is accurate and affordable, build it into daily work, and expand once it pays off, said Mr Chanwit. Thai developers can add value through localisation, integration and ongoing service.

Each business should pick one task, record a baseline, name a manager and run a trial of 60-90 days where data and integration allow, he noted. Useful measures include processing time, defect rates, energy use and customer response times.

Matching grants or vouchers should depend on independent proof of business value, with a clear budget and owner for scaling up. Workshops should lead to real workplace projects.

"Government procurement can give qualified Thai AI firms a first customer through smaller, transparent contracts with clear rules on data protection, reliability and support," Mr Chanwit said.

KEY CHALLENGES

He said the biggest hurdles are proving value, preparing data and finding people to put AI to work.

Even low-cost AI involves real spending on cleaning records, linking older systems, checking outputs and maintenance.

"We need reliable Thai-language and sector-specific testing, affordable support beyond Bangkok, and clear rules on who controls data and handles failures," said Mr Chanwit.

Smaller firms need practical privacy and security guidance. Staff must understand how their roles will change and have time to learn, he noted.

Manufacturing and healthcare should lead Thailand's next wave of small AI adoption, said Mr Chanwit.

Food processors, electronics makers and industrial suppliers have repetitive tasks where AI can help with visual inspection, maintenance and workplace safety. Defects and downtime offer clear baselines for measuring results.

In healthcare, early uses include appointment and queue management, clinician-reviewed documentation and forecasting demand for medicines and supplies. Success can be measured by staff workload and patient waiting times, he noted.

Tourism, hospitality and retail are also well-suited to wider SME adoption of small AI, from guest inquiries to paperwork and stock forecasting.

Agriculture has strong potential too, especially through cooperatives and farm service providers that offer trusted advice and local support, Mr Chanwit said.

Small AI can be attacked through its software, connected devices, data, or the model itself. Being small or on-site does not automatically make it secure.

Risks include stolen passwords, leaked customer data, tampered model downloads and corrupted training data. With language-based tools, attackers can hide malicious instructions in documents or messages, an attack termed a prompt.

SMEs can use AI faster than large companies because they have fewer legacy systems and entrenched processes, said Axel Winter, chief executive of Xponential, a technology and AI service firm.

Rather than seeing limited technology resources as a disadvantage, SMEs should build AI skills across their workforce and encourage experimentation, he said.

JOB DISPLACEMENT

Regarding the economic risks of AI for Thais, Putthiphan Hirunyatrakul, research fellow at Thailand Development Research Institute (TDRI), said an analysis of online job ads found the share mentioning AI skills rose 3.9-fold between the second quarters of 2024 and 2026. Demand extends across a range of occupations, beyond IT.

AI is already taking over some tasks. Occupations expected to face the greatest impact include office clerks, data entry, accountants, translators and marketers, along with workers in creative industries such as advertising, where AI can handle everything from audio production and photography to editing.

The impact is already apparent among recent graduates, who face greater difficulty finding work as AI takes over many entry-level tasks, he said.

TDRI estimates 2.2 million Thai workers are at risk of AI displacement, representing the third-highest share of the workforce in Southeast Asia after Singapore and the Philippines.

One concern is that well-paid jobs offering strong career prospects may be among the first to be replaced. Meanwhile, occupations that remain harder to automate, such as skilled trades, wellness services and elderly care, tend to offer lower pay.

Mr Putthiphan said the government should help businesses adopt AI to improve productivity, while also encouraging the redesign of jobs and the creation of new career pathways for affected workers.

Mr Chanwit said small AI can replace tasks and sometimes whole jobs, especially routine administration, data entry and standard customer inquiries. It would be misleading to promise that every worker will benefit automatically, he noted.

AI can also ease Thailand's labour shortages, while lifting output and growth. An ageing society will not stop some jobs from disappearing, and young people need first jobs to gain experience. The net effect on jobs is unclear, said Mr Chanwit.

"Employers should train staff early and redesign roles around judgement, customer relationships, technical expertise and accountability," he said.

Meanwhile, the government should back paid retraining, apprenticeships and income support during the transition, said Mr Chanwit.

"I believe the success of AI should be judged partly by better work and higher earnings. We should track job impacts alongside productivity and act when some groups gain while others bear the costs," he noted.

NEEDED SKILLS

AI is reshaping the skills that Thai employers seek more than creating a standard set of new job titles, according to Duangporn Prom-on, managing director of recruitment platform Jobsdb by SEEK Thailand.

Demand is growing for professionals who can apply AI across data analytics, software development, digital products, customer experience, cybersecurity and digital marketing.

More employers want domain expertise paired with practical AI skills: working with data, using tools responsibly, critically assessing outputs and protecting confidential information.

As AI automates or assists with routine tasks, including administration and documentation, Ms Duangporn said analytical thinking, creativity, communication, judgement and problem-solving become more valuable.

ACCESSIBLE FOR ALL

Payong Srivanich, chairman of the Thai Bankers' Association (TBA), said making AI accessible to everyone is essential, particularly in helping SMEs improve their livelihoods and gain better access to financing.

"What matters is how AI can help people across all segments of society, particularly small traders and street vendors, improve their livelihoods and build on their existing work to gain better access to finance," said Mr Payong, also president and chief executive of Krungthai Bank.

AI can help small traders overcome administrative barriers to borrowing, as many lack the time and resources to prepare financial statements or collect the bank statements and other documents typically required by financial institutions, he noted.

AI tools could make it easier for small businesses to demonstrate their financial standing and apply for loans.

This effort aligns with the government's 60:40 "Thais Help Thais Plus" co-payment scheme that aims to ease the cost of living, as it provides an opportunity to introduce these tools to small vendors, said Mr Payong.

The initiative complements the Finance Ministry's efforts to develop alternative data for credit scoring, as well as a data exchange bureau being advanced jointly by the ministry, the Bank of Thailand and TBA, he noted.

These projects promote financial inclusion while addressing the risks of financial exclusion and cyberthreats associated with the growing use of digital payments and AI tools, said Mr Payong.

"Expanding access to finance must go hand-in-hand with measures to protect users, as digitalisation also brings new forms of cyber-risk," he said.

The high cost of advanced technologies often prevents smaller businesses from benefiting from developments in AI, with attention frequently focused on large-scale AI systems and large language models that are expensive and smaller operators may struggle to access.

Thailand is working on several initiatives to address this gap, including Ari Score, a proposed data exchange bureau, and the development of prompt lists to help users apply AI more effectively. Another proposal promotes government procurement of products and services made in Thailand by Thai businesses.

Ari Score is a new credit-scoring initiative by the Finance Ministry that uses big data and AI to assess borrowers' creditworthiness, seeking to expand formal loan access for SMEs, street vendors, small traders and individuals who lack a credit history with the National Credit Bureau, or have previously been blacklisted.

"Taken together, these initiatives could help create a more integrated economic system in which reliable data supports lending by banks and enables private companies to develop new business opportunities, particularly for smaller enterprises," Mr Payong said.

The government also needs to provide appropriate incentives and review regulations to establish a level playing field, he said.

"Technology can improve the efficiency of the economic system, but we need to ensure that all sectors can benefit from it, particularly smaller businesses," said Mr Payong.

Thailand must accelerate implementation and coordinate efforts across the public and private sectors to keep pace with global and regional changes, maximising the benefits of technological transformation for the country, he noted.

USAGE BY SMEs

AI is expected to enhance the competitiveness and productivity of SMEs without causing significant disruption to employment, said Pimjai Leeissaranukul, chairwoman of the Federation of Thai Industries (FTI).

"AI serves as a tool to improve efficiency and reduce workloads, not to replace workers," she said, adding the FTI remains committed to helping SMEs adopt AI technologies.

Most SMEs operate with relatively small workforces, meaning AI is unlikely to lead to widespread job displacement, said Mrs Pimjai.

Instead, the technology can take over routine administrative tasks and daily accounting work, allowing employees to focus on higher-value activities.

Business operators can play an active role in developing and applying AI solutions to improve productivity in areas such as quality control, production planning and energy management, she noted.

Entrepreneurs should invest in training employees so they can work effectively alongside emerging technologies, said Mrs Pimjai.

To support this transition, the FTI partnered with the Office of Small and Medium Enterprises Promotion to launch the "Smart AI for SMEs" programme, backed by 85 million baht in funding.

The initiative aims to assist 3,000 SME operators across all industrial sectors, addressing a wide range of business needs, including accounting, sales, manufacturing and business planning.

The programme is intended to strengthen the competitiveness of Thai enterprises in a digital economy.

Project preparations are focused on personnel development and gathering feedback from entrepreneurs who expressed an interest in participating, said Mrs Pimjai.

The government approved the budget in August, and disbursement is expected later this year.

"Once the funding is released, we will immediately move forward with implementation, including technology deployment, training programmes and hands-on pilot projects with participating businesses," she said.

HIGH IMPACT

Maruay Tangmitrphracha, executive director of SET-listed Dohome, a retailer and wholesaler of construction materials and home renovation equipment, said integrating AI into operational workflows requires relatively low investment but can deliver high-impact results.

The company has continuously embraced new technologies, integrating AI across multiple business functions.

Regarding customer service, he said he believes leveraging AI enables the company to meet customer expectations and fulfil needs more accurately than businesses who have yet to adopt it.

A key application of AI at Dohome involves generating predictive recommendations based on customers' previous inquiries, purchase histories, and predicted interest in related product categories.

"Customers may not focus on the technology behind the scenes. What matters most is whether the service provider can recommend products that directly solve their problems," said Mr Maruay.

The company already deploys an AI-powered chatbot, reducing the contact centre workload by up to 70%.

Two years ago, Dohome introduced an AI-driven demand forecasting system to dynamically manage inventory levels, significantly reducing overstock.

"For the company, AI adoption is still in its early stages," he said.

Store sales staff are being trained to retrieve basic warehouse data and use AI-powered tools to answer routine customer inquiries.

The goal is for frontline staff to become technologically proficient, using AI tools in their daily routines, said Mr Maruay.

About four months ago, the company organised an internal hackathon, inviting teams from various departments to compete in solving workflow bottlenecks using AI.

The winning project used AI to aggregate scattered departmental data across sub-branches, analyse and prioritise branch-specific issues such as customer service and inventory management, and provide recommendations to managers.