Exclusive: ‘Putting the AI in Thai’ – Sam Tanskul, Krungsri Bank in “The Fintech Magazine”
Krungsri Bank made a shrewd decision four years ago to get to the head of the queue when it came to acquiring and investing in new technology, by creating Krungsri Finnovate. Its MD, Sam Tanskul, believes it can now help propel Thailand towards an AI-enabled future.
It’s no coincidence that the biggest second-quarter financial deal in Southeast Asia involved Thai bank Krungsri and its Japanese superpower parent Mitsubishi UFJ Financial Group (MUFG).
The $706million investment in Singapore-based Grab Holdings, the ride-hailing-turned-super app, which now includes food delivery and financial services, was a shrewd one from a shareholder’s perspective. Grab is, after all, South East Asia’s first decacom, valued at more than $10billion. But Krungsri is more interested in what it can learn from Grab’s mission to become an ‘AI everywhere operation’.
To put this in perspective, every day Grab generates about 40 terabytes of data, which is analysed using artificial intelligence (AI) to bring improvements for customers and the nations it operates in. And it’s in no mood to stand still, with its AI development budget, encompassing infrastructure, talent, technology and strategic partnerships, totalling more than $100million in 2019 and expected to increase by 50 per cent in 2020.
The vital importance of AI to future prosperity is well understood by Krungsri (the common name for Bank of Ayudhya, Thailand’s fourth largest bank), and the country as a whole. Thailand has put AI at the centre of national policy with its Digital Economy and Society Development Plan, and is testing a biometrics-based digital identity scheme for all citizens which will be a foundation stone for the digitisation of infrastructure that touches every aspect of their lives.
Based on blockchain, the biometrically-supercharged Smart ID system will provide near-instant, secure verification of individuals and a gateway to the future for Thais, be it remotely opening a bank account using e-KYC (know your customer) or accessing medical help through telemedicine. Indeed, according to a Thai government-commissioned survey by The Economist Intelligence Unit (EIU), ‘without an effective industrial policy that harnesses AI, Thailand risks losing six per cent of GDP (gross domestic product) in 2035 due to losses in export competitiveness and declines in foreign direct investment’. But by introducing the right interventions to harness AI and associated automation tools to boost skills, strengthen international competitiveness and improve public services, ‘Thailand could keep unemployment levels low and see an increase in GDP to more than US$1trilllion by 2035, up from US$544billion today’.
For its part, Krungsri recognised the value of technology investment back in 2016, when it set up Krungsri Finnovate, a venture capital fund for fintechs which also runs its fintech accelerator programme.
It has a two-pronged strategy: to use the accelerator programme, Krungsri Rise, and a graduate contest called Krungsri Uni Startup, to expand Thailand’s fintech ecosystem; and a $30million fund aimed at growth startups at Series A funding stage or later to make sure they succeed. `It’s aiming to be the best banking corporate venture capital business in ASEAN (the association of Southeast Asian Nations) by channelling investment specifically into AI and data analytics innovation.
Sam Tanskul, MD of Krungsri Finnovate, says its RISE project was developed from a model successfully used by its parent bank MUFG. Collaboration is at its core.
“The first year was 2016 and, at that time, there were only 20 fintechs in Thailand, so we invited everyone to join our accelerator, in order to learn and see how the fintech startups work,” says Tanskul. “The main objective of Krungsri RISE is to create a fintech ecosystem in Thailand.”
Now entering its fourth year of running the accelerator, the Krungsri Finnovate fund has invested in the winning fintechs from each of the previous RISE rounds. The fintechs spent three months on campus and had to prove that they could work with the bank’s business units on demo day.
“They have been used from that year until now,” says Tanskul. “We believe that fintechs will be successful if they have synergy projects with the bank, and we are open to working with them.”
Among its AI investment successes so far are wealth management platform Finnomena, and Baania, which uses machine learning of real estate and consumer housing data to improve efficiencies in buying, selling, investing or building real-estate projects.
In January 2020, Krungsri Finnovate was also among a group of investors which helped Finnomena, which has rapidly become one of the biggest wealth management platforms in Southeast Asia, raise a further $10million in a Series B funding round.
But perhaps a more significant investment was the one it made in 2018, when Krungsri Finnovate went beyond the Thai border to become part of leading Japanese fintech and venture capital firm SBI Holdings’ AI & Blockchain Fund. That started it on a journey to help identify global startup players to further its mission to leverage the best AI and blockchain technologies and partnerships that benefit the bank’s own operations and achieve the imperative to cut operating costs while improving customer satisfaction.
The move led to it being be part of an SBI-led consortium of investors who, in March last year, ploughed $8million in Series A funding into Silot, a Singaporean fintech company that helps banks improve their decision-making processes using AI.
Tanskul points to what he perceives to be a different attitude to partnership working in the ASEAN region compared to the UK and Europe.
“In the UK or Europe, most of the fintechs are banks’ competitors,” he says. “I think we have a different business objective. In Southeast Asia, currently, most of the startups are not the banks’ competitors; most of the fintech startups are banks’ companions. The fintechs have the technology, banks have the customers, so we look at how can we collaborate with each other, not compete.”
One of the bank’s objectives in using AI is to create what Tanskul calls hyper-personalisation. “More than personalised, it means we have to know and understand customers, offer the right products, at the right time, in the right place. This is the motto of the bank,” he says.
“Most of the big banks in Thailand have a ‘super app’ mobile banking application. We are not only offering banking products to customers: you can make a hotel booking on our mobile app, or book an Uber or Grab, because we want to capture the data around customer lifestyle. We need to understand more, if we are to present the right product to the right customer.”
There is a strict development structure in place within Krungsri to ensure that the developers don’t produce something the bank can’t work with.
Tanskul explains: “The external fintech team source the technology, my innovation lab work with the new technology, either from a startup or bigger tech company, running proofs of concepts to test whether this technology is matched to a bank’s business unit. If it tests well and if we want to roll it out, then we remove it from the innovation lab team and give it to the business unit, where we have another team working to run the project with them.
“In that way, we became the first bank in Thailand to launch a chatbot, with voice and non-voice. Right now we are working on a project called AI Voice with a big tech team from Asia and a local Thai team. Our innovation lab tested whether the voice – Thai voice and Thai language – using natural language understanding (NLU) was acceptable for our banking product. We tested it for a year and it was proven to work, so then we handed it to our call centre team, to roll out. So this has been end-to-end.”
Banks in Thailand have no alternative but to embrace the most advanced digital ways of working.
“The interest rates and fees have been capped by regulation, so we have to see how we can reduce operational expenses,” says Tanskul. “We’ve found that digital products are key. Right now, every sale is online, we don’t print documents, and things that might have taken seven days and 10 staff can now be done within half an hour.”
When it comes to eKYC, for which most Thai banks have, for the most part, adopted Chinese facial recognition technology, Tanskul says, they are also under pressure. “The government of Thailand has said it doesn’t want any fraudulent cases of e-KYC – not a single person,” he says.
For that reason, banks came together under a National ID initiative to use blockchain to automatically cross-reference the identity of a new customer. As Thailand heads towards an AI-enabled future, Krungsri won’t be alone in using the most intelligent technology… but it hopes to lead rather than follow.