Kasikorn Business-Technology Group (KBTG) has declared 2024 as the year of turning technologies into products. KBTG has many researchers and tech developers working devotedly on deep tech with business potential and promising user benefits. The company is aware that such advanced deep-tech research will be meaningless unless applied with a “human-first” goal to benefit the masses.
KBTG Chairman Ruangroj Poonpol said that the company in 2024 aims to drive technological development continuously with machine learning, artificial intelligence (AI), and deep tech – or M.A.D. in short. The goal is to move from “AI first” to “human first” to make actual use of the technologies in this group.
“KBTG’s goal for 2024 is to apply the research works so that they bring benefits to the masses. Also, we will share our knowledge and direct experience to benefit business units of Kasikorn Bank, KBTG, our partners, and the country,” Ruangroj said.
In the past, KBTG was driven by the concept of being an “AI-first organization” to become a Thai leader in M.A.D. Regarding AI, the focus is on research through KBTG Labs and important allies like the Massachusetts Institute of Technology (MIT), and on application through Kasikorn Bank. Several actual products and services have been created for actual use, and many more in the pipeline are expected to launch in the market in 2024.
KBTG Labs innovations for businesses
The deep tech research team of KBTG Labs has turned their findings into many actual products. Much of the work involves innovations beneficial to businesses, such as verification technology called “AINU”, “Car AI” damage examination system based on AI technology, and intelligent banking analysis system.
– AINU: AI technology is used for this verification/identification service involving an optical character reader (OCR) to extract data from ID cards. The service is aimed at preventing identity theft through the liveness detection technique and deterring potential fraud with face recognition technology. AINU, already in service, offers SDK and API technologies to digital banking customers.
– Car AI: A collaboration with Muang Thai Insurance, this system allows customers to report car accidents for insurance claims without the help of a claim official. The user can take photos of their cars involved in an accident and use Car AI to estimate the damage. KBTG won the Asian Technology Excellence Award 2023 in the category of AI – Financial Technology from the Asian Business Review, a leading business magazine in the region.
– Intelligent banking analysis system: KBTGuses M.A.D. technologies and research work to create this intelligent system for banks. It is used to complement Kasikorn Bank’s various products and services, including Marketing Intelligence, Credit Intelligence, Fraud Intelligence, and Call Center. The technologies used in data evaluation and analysis are worth an estimated 500 million baht.
KBTG places much importance on conducting research and applying the study findings for actual benefits. The company aims to enhance the potential of M.A.D. technologies for businesses and industries by carrying out in-depth research and developing new products based on the technologies to truly meet market demand. Commercial services are provided through KX, which is open to financial partners and businesses from other industries.
More work in the pipeline
– CCTV Analytics technology: Data in the form of video images from close-circuit cameras is analyzed by AI to determine the behaviors of people in the video recording. Experiments have been carried out at certain Kasikorn Bank branches to count people and observe people’s movements for analysis and improvement of the branches’ areas and services involving tellers at the counter or self-service by customers. This technology can be applied to any other business that requires similar analysis, such as food shops.
– Smart check-out technology: This is intended for shopping malls and supermarkets. AI is used to analyze images of goods and calculate the prices from the database. When goods are placed on the check-out counter, the camera takes their photos as AI calculates the total prices based on the database. This technology is still in the stage of “proof of concept” (PoC), which means the current work is focused on determining whether the idea can be turned into a reality. More study and development are needed before it can be marketed. The current work still cannot differentiate between goods in cans of different sizes, as the camera “sees” that containers with a round lid are of the same size.
– Wealth management technology: Machine learning AI is used to provide advice on stocks and funds to customers. This work is still in the development stage at KBTG Labs, using information on customers, stocks and funds from 8-10 years ago. Formulas for investment and portfolio management are applied in the calculation to get investment advice for customers. Macroeconomic data and current news will also be used to train the AI. This service is expected to be available by late 2024.
– Thai NLP: A chatbot called “Nong Ruang Khao” is created with generative AI. Its advantage is that generative AI allows the chatbot to respond to user questions better than the current version employed by K Bank Live, which often gives long, uniform answers. “Nong Ruang Khao” can chat with customers more naturally as if they were talking with humans, making it more likely for customers to prefer interaction with the chatbot rather than phoning the Call Center. “Nong Ruang Khao” can offer general advice involving financial matters, credit cards, and basic loan evaluations, among others. The chatbot is still in the development stage.
– VDO Analytics system: AI is used to analyze video recordings and provide such information as how many people are involved in the video, who said what, what kind of feeling each speaker had, and whether there was any expletive in their conversation. The recorded conversation is transcribed, and each sentence is analyzed to determine the sentiment of what was said in the video. This system is not available commercially.
International quality recognition
The quality of KBTG’s research work on M.A.D. is recognized internationally, as KBTG Labs has a large group of AI researchers and developers. They are among the largest groups in Thailand working on various areas, including facial recognition, computer vision, natural language processing (NLP), and deep learning. The research is carried out by the company’s researchers and developers to ensure quality for future commercial use by Kasikorn Bank, KBTG, and their business allies.
In 2023, KBTG Labs had its research studies on M.A.D. published in six international publications. Also, it sponsored a research project of KBTG fellow Pat Pataranutaporn, the first Thai doctorate candidate at MIT Media Lab, a research laboratory at the Massachusetts Institute of Technology. The topic of his research is “Influencing human-AI interaction by priming beliefs about AI can increase perceived trustworthiness, empathy and effectiveness”, which was published in Nature Machine Intelligence, a scientific journal covering machine learning and artificial intelligence.
KBTG Labs collaborates with MIT Media Lab
KBTG Labs has joined hands with MIT Media Lab, a world-class research institute, exploring new technologies. Among their collaboration is “Future You” (an AI platform that helps users to chat in real-time with their “digital twin” in the future) and “Kookid” (a digital “consulting partner” that functions as a chatbot for counseling, with information from reliable sources).
Two conversational agents – Kana” and “Kacha” – are used to offer opinions from different perspectives for maximum efficiency in decision-making. Kookid is developed from Knowledge-GPT (K-GPT) research work that is based on ChatGPT capabilities in offering in-depth knowledge. It also is capable of understanding the Thai language. There is a plan to launch Kookid in the market in 2024.
Using and sharing M.A.D.
KBTG adopts technologies involving machine learning, AI, and deep tech not only to develop products and services for other organizations but also to improve the work efficiency of KBTG developers. For example, Copilot is used to help improve the speed of coding by two times. The M.A.D. Guild was set up in the organization to serve as a platform for sharing knowledge and allowing easy access to AI for KBTG staff members.
Moreover, KBTG aims to play a role in improving Thailand’s AI literacy by bringing to a public ecosystem its knowledge about machine learning, AI, and deep tech earned from research and internal skill sharing.
KBTG is determined to share knowledge internally and externally to help improve Thailand’s capabilities on M.A.D. The KBTG Kampus ClassNest offers M.A.D. Bootcamp courses for IT personnel looking to improve their M.A.D. skills.
The company also encourages the AI community to organize activities related to M.A.D., as well as sharing best practices and lessons learned from KBTG’s direct work experiences through KBTG Techtopia and other knowledge-sharing forums.
KXVC investment in deep tech and AI
In addition to its research work through KBTG Labs, KBTG has also launched KXVC, a venture capital of 3.5 billion baht that invests in AI, Web3, and deep tech.
This is due to the company’s confidence in the significant roles of AI and deep tech in the business world. KBTG is aware that venture building alone is not enough, so it invests to help expand the potential of development, both in terms of quantity and quality. That led to the formation of the KXVC fund with an initial investment of about 15 million US dollars.
In 2024, KXVC is prepared to go full steam for investment in AI startups and funds around the world, serving as a gateway to bring new AI technologies and overseas investments into Thailand.
KBTG is ready to work with its partners in Thailand and other countries to improve M.A.D. capabilities and turn its research work into products for the masses.
All those works and plans confirm KBTG’s ambition to become an M.A.D. leader in Thailand, which is part of its strategy to be an AI-first company that uses deep tech to create human-first services for the benefit of people and the country.
