SCBX, a leading financial technology group in Thailand, has launched its inaugural report titled "SCBX AI Outlook 2025: Beaconing the Future of Artificial Intelligence," marking a significant milestone in the country's digital landscape. This report outlines four key trends that are set to transform Thai businesses as they adopt an AI-first approach, with the ambitious goal of generating 75% of its revenue from AI by the year 2027.
Kweewut Temphuwapat, Chief Innovation Officer of SCB X Public Company Limited (SCBX), emphasized the company's commitment to becoming a fully AI-driven organization. "We are determined to create 75% of our revenue through AI technology by 2027, and we are systematically developing advanced AI capabilities through the use of complex data and targeted R&D initiatives," he stated. The report serves not only as a forecast of technological trends but also as a strategic roadmap for organizations aiming to transition into AI-first entities.
The SCBX R&D Innovation Lab has been established as a central unit for research and development of AI technologies. This lab collaborates closely with SCB 10X and DataX to leverage AI innovations, aiming to create impactful prototypes for practical applications. The Typhoon Thai LLM (Language Learning Model) project exemplifies this effort, showcasing how SCBX is integrating AI into various sectors to enhance operational efficiency and decision-making processes.
In the report, SCBX identifies four main trends that are expected to reshape the business and industrial landscape:
1. Open vs. Closed AI Models: The report highlights the intensifying competition between open-source and closed-source AI models. In the early days of Generative AI, the landscape was dominated by closed models from major players like OpenAI and Google, which treated data as proprietary and restricted access. However, open-source models like DeepSeek R1 are gaining traction, allowing developers to innovate rapidly and at lower costs. SCBX references a Stanford HAI study indicating that in 2023, over 65% of new AI models released were open-source, a significant increase from 33% in 2021. This shift signals a decentralization of power in the AI realm.
2. The Rise of Tiny Titans: Another trend noted in the report is the emergence of smaller, more efficient AI models that prioritize performance over size. This includes multimodal models that can process various data formats simultaneously—text, images, and audio—mimicking human sensory perception. Small Language Models (SLMs) are being developed for specific tasks, reducing resource consumption and enhancing privacy by processing data on-device rather than on remote servers.
3. Enhancing Reasoning Abilities: The report also discusses how improving AI's reasoning capabilities does not always require larger models. Innovations in model architecture, training techniques, and data design can enhance performance significantly. A case in point is Microsoft's Phi-4, which, despite having only 14 billion parameters, outperforms larger models like Gemini Pro 1.5 in complex reasoning tasks.
4. Agentic AI: The concept of Agentic AI, which can plan, decide, and act independently, is highlighted as a pivotal development in the AI landscape. Dr. Tutanun Sinthuprasirt, Head of Emerging Tech and Venture Building at SCBX, explained that Agentic AI operates similarly to humans by learning from experiences and adapting to complex processes. This advancement allows organizations to focus on strategic thinking rather than merely managing AI systems.
SCBX's report also raises critical questions about the future of Artificial General Intelligence (AGI). It notes that there is no definitive answer yet on whether current language models are sufficient for AGI or if new architectures are needed. Experts like Sam Altman, CEO of OpenAI, assert that language models are central to achieving AGI, while others, including Mustafa Suleyman of Inflection AI, argue that AGI will require fundamentally new structures that mimic human brain architecture.
Furthermore, the report emphasizes the importance of open-source technologies in fostering innovation and reducing costs. For instance, the integration of DeepSeek R1 with SCBX's Typhoon model aims to enhance Thai language comprehension and logical reasoning capabilities, allowing the organization to gain a competitive edge in the market.
SCBX is not merely theorizing about AI; it is actively applying these technologies across various sectors. For example, the Typhoon model is being utilized at Siriraj Hospital to improve knowledge management and reduce the workload of healthcare personnel. Additionally, the company is developing legal chatbots that provide timely and accurate responses across 35 areas of Thai law, as well as analyzing labor market trends in collaboration with the Thailand Development Research Institute (TDRI).
As SCBX moves forward, it encourages other organizations to embrace open-source AI and develop Agentic AI systems that enhance decision-making capabilities. The report concludes with a call to action, urging companies to rethink their strategies regarding AI, asking not how to use AI, but how to create new business models with it.
In conclusion, SCBX's AI Outlook 2025 report presents a comprehensive and strategic vision for the future of AI in Thailand, emphasizing the need for organizations to adapt to the rapidly evolving technological landscape. With its ambitious goals and innovative initiatives, SCBX is positioning itself at the forefront of the AI revolution, paving the way for a more efficient and intelligent future in business.
