How Bangkok Bank worked with Pand.ai to develop a conversational AI engine to better service customers

Conversational Artificial Intelligence (AI) is the technology that allows chatbots to respond to natural languages, similar to a human. However, unlike the latter, a chatbot never gets tired and can provide round-the-clock support with fewer errors.

A chatbot powered by advanced conversational AI delivers a better customer experience by helping businesses increase customer engagement and improve operational efficiency.

Bangkok Bank’s innovation programme InnoHub and Singapore-based fintech startup Pand.ai have joined hands to develop TT01, a Thai conversational AI engine. It is the first time such an engine has been built in-house by a Thailand bank.

Development began earlier this year and the engine was fully developed in October. Bangkok Bank will use the engine as a digital sales assistant for its sales and relationship managers. The chatbot is scheduled to launch in Q1 2021.

Pao Sriprasertsuk, Head of Bangkok Bank Innovation Department, and Shin Wee Chuang, Co-founder and CEO of Pand.ai shared in an interview with e27 their take on topics, including developing a Thai conversational AI engine, the keys to ensuring the older generation can benefit from new technologies and future digitalisation trends in the banking industry among others.

Below are the edited excerpts from the interview:

What were the key challenges faced when creating the AI engine and how did the team overcome them?

Chuang: The biggest challenge we faced was that our data scientists were not native Thai speakers. When it came to building a Thai AI engine, this resulted in a loss of efficiency.

The data scientists had to refer back to the Thai speakers in the team to check if everything is correct. Hence, it was not as intuitive and efficient as developing an AI engine in their native language.

Also Read: Artificial intelligence is a key consideration for companies looking to adapt operations to optimise user experience

Since we were collaborating with Bangkok Bank, we had two sets of data scientists, one each from Bangkok Bank and Pand.ai. Therefore, we split the workload; the processes that required the knowledge of the Thai language were handled by the Bangkok Bank team, while we handled those that did not depend as much on the ability to understand the Thai language.

It was a collaborative effort for which both sides pitched in with what they were good at.

Why are there a lack of Thai chatbots in the market?

Chuang: One of the primary factors, I believe, is the lack of market opportunities. There are a lot of people speaking Thai and Thailand itself accounts for close to 70 million people.

However, in the grand scheme of things, when compared to the number of people speaking English, Mandarin or Spanish, the number of people conversing in Thai pales in comparison to these major languages.

That would that limit the commercial appeal of developing a Thai conversational AI engine.

Secondly, AI is a big field, and natural language processing (NLP), which powers conversational AI, is a subsegment of it. Not every data scientist’s speciality is in NLP.

They could specialise in computer vision and be largely clueless about NLP. Therefore, among the data science community, the number working on NLP is small and it is even rarer to find a Thai native speaker among them.

Does digitalisation represent a “must-have” rather than a “nice to have” for banks today?

Sriprasertsuk: Digitalisation has become a must-have, as it drives key competencies for banks in this digital age. Banks are now expected to provide ‘good’ digital services.

Definition of ‘good’ may vary by individual. To some, it may mean ease of usage and real-time support. To others, it might be tailored recommendations and security features.

Ultimately, these services can be only driven by digital technologies such as automation, data analytics, and AI. They also reduce unit costs while allowing banks to serve customers at scale in a wider, faster, and cheaper way.

Also Read: 4 ways the banking sector can respond to the digital transformation

However, that does not mean that traditional relationship-based banking or physical bank branches will disappear completely.

We do believe that human elements, such as relationship and offline services, will still play an important role as one of the key competitive advantages for banks.

How do you ensure a diverse customer base (especially the older generation) can keep up with the rapid introduction of technological solutions?

Sriprasertsuk: Empathy is key. Different customer segments have different needs and these needs are ever-changing.

For the older generation, learning new technologies remains one of the biggest hurdles for them.

However, it may be surprising to find that they manage to use ‘Paotang’, an e-wallet where Thai government provides government subsidies to the public in the form of e-money in order to boost the economic recovery during COVID-19.

The seniors were able to use the e-wallet due to the help of their families, friends and merchants, who helped guide them through the process, and they mastered it after a few tryouts.

Therefore, support from the people around seniors is a helpful enabler for them to learn and adopt new technologies.

Driven by customer research and Big Data analytics, banks will also need to be empathetic and design solutions that are catered to the needs of their customers.

Take the chatbot as an example, where customers may struggle to get the right keyword to ask a chatbot.

By adding human elements to the chatbot such as programming to analyse the intention of a user or remembering the context of a conversation, it can make an educated response to a user’s inquiry and create a better customer experience.

In what areas of banking do you see technology has the greatest impact in the near future?

Sriprasertsuk: AI and Big Data analytics are technologies that would provide big leaps to this digital era for all industries, not only the banking sectors.

For banking, AI has transformed every aspect of bank function, allowing banks to deliver personalised digital banking experiences, which are seamless across touch-points. AI can also enhance the efficiency of banking processes, reduce costs, improve security and strengthen risk management.

Also Read: Reimagining anti-money laundering processes with blockchain technology

Distributed ledger technology (DLT), more commonly known as the blockchain technology, is expected to play an important role in the future of financial services, although it still needs improvement.

DLT helps enhance security, traceability, efficiency and speed of transactions. The technology allows banks to collaborate better with stakeholders in trade flow, from improving traceability to greatly reducing time and costs.

Additionally, DLT also enables the concept of smart contract and programmable money that could lead to a wide variety of new digital financial products.

The post How Bangkok Bank worked with Pand.ai to develop a conversational AI engine to better service customers appeared first on e27.