DBS Digital Transformation: AI, Blockchain, and Customer-Centric Banking

up next may I please invite on stage tan suan Deputy CEO and head of institutional banking DBS

group well let's welcome her with a huge round of applause ladies and gentlemen Miss tan sushan thank you so

much thank you I just arrived this morning from Singapore and um just this week a lot of

uh Singaporean ministers and um and Indian ministers were talking a big topic was digital payments a big topic

was how the two countries can work together better so today I'm here to share with

you DBS for those of you you don't know DBS is the Development Bank of

Singapore we call ourselves DBS now because we're now a Commercial Bank a big Asian Commercial Bank the biggest in

Southeast Asia um but I'm here to share with you our digital Journey how it began what we had to do after all banks

you know are not like finex banks are big who are very heavily regulated we serve you know millions of people and

um it's not easy to move so today speaking at the global fintech Festival uh thank you to to to to NPC and and and

and and the sponsors I'm here to share with you how we did this transformation inside

out so firstly it begins with the body and what is the body the body is the technological framework

second it's the mind what is the mind it's basically the people in the organization and how we've changed as an

organization in the way we treat our customers putting the customer bang in the center of how we look after them and

thirdly the soul and what is the soul the soul of a company is essentially its culture it's the business business

practices it's the habits it's what binds the employees together as a company with a common purpose a common

vision and a common Behavior so let's unpeel this so being Tech being technological driven what

does that mean for an organization most banks used to have big monolithic systems right and a lot of

them used to out Outsource it to Big tech companies we were no different and what

we did very early on was we looked at all the various systems we had and we realized what is banking banking is

actually people and Technology right it's HR and Tech my my my CEO P likes to call it hit HR

it and that's what it is is bringing the best of the people combining it with the best of your technology and giving your

customer the best experience but if you Outsource all of it how do you control it so we

decided we should really insource a lot of our Tech we used to be 85% outsourc we're now more than 85%

insource and then we looked at our Tech stack from monolithic we went to microservices like all good Tech

startups like all good fintech companies going straight to the cloud but also understanding in a balance between going

to the public Cloud being agile and quick and balancing that with a resiliency that comes from um uh being a

big bank and and also the resiliency of of of frankly the big monolithic systems and so having that hybrid we came up

with a hybrid Cloud solution we call it our hybrid private Cloud where we still use the best of the compute of the

public Cloud but we have a paved gateway to make it get secure for the bank and then of course API

everything in this respect we were lucky we came up with our API suite and our API protocol we've got some several

thousand apis and we came up with it very early in 2016 so that by the time covid hit guess

what we were ready because in 2019 I told my team hey let's have this API war cry let's get as many of our

clients as possible as many of our corporate clients big and small onto our apis our apis to look at balance to look

at transactions to make transactions to make trade you know inter inter country trade um and to do any kind of

transactions so when we came up with our API Suite we called it realtime API by DBS rapid for short and really getting

our CTO to talk to the client CTO and getting our API Suite to to customers and enabling them so for

example insurance companies right when you miss your flight normally well you don't if if your flight is delayed

normally it's a slow form filling exercise to show your ticket show that the flight was delayed but if you've got

a realtime connection then if the flight's delayed more than 4 hours basically the customer of the insurance

company company can immediately send through an API call a claim and we can verify all the all the information and

send the claim back immediately instant claims in India for example we've enabled logistics company truck drivers

Etc to track the the the truck drivers where they're going help them to pay for the petrol get them paid per per per per

trip cashless data driven and I think this has it's been the big game changer for

us and then getting Tech Talent you know we have 12,000 Engineers working at DBS and we have less Bankers we have

probably like 8,000 Bankers so we like to say that we're we're a bank but we are really a TCH enabled Bank we've got

more Engineers than we do bankers and frankly that's the way the business is going and um to be be

honest a bit of gender imbalance in the past especially in in in the stem front and so we started to have hackathons for

women only women Engineers so women in tech hackathons for women um and a lot of dedicated campaigns to try and hire

through AI gen and um and and and you know crowdsourcing great ideas from the front looking forward for one are the

few things that we're looking at to get still in the Forefront of technology and here the two buckets one the Hot Topic

generative Ai and the second is the blockchain so let me unpeel this I don't know if any of you read DFT yesterday

but the head of Microsoft the CTO of Microsoft said they would stop hiring Engineers that don't compute with AI by

the end of this year I thought that was quite a seminal statement to make and um I wanted to talk I wanted to UNP this

generative AI frankly AI is not new to any of us right everyone's been using AI for years

at DBS we started this AI movement probably 2016 17 actually no we started this in

2014 I was running wealth management my team came up to me and said let's use AI to help our wealth

Bankers uh you know be more contextual in their advice to clients and I said okay but the 2014

large language models were not as developed and so there was a lot of false positives and frankly we had to do

too much work to teach the machine you know how to read how to write what to read what was what was more important

was it the time was it what the FED said what the boj said it was very difficult so that was our first experiment and

frankly it failed but fast forward now with llms where they are and with generative AI

which I think is an incredible because a generative AI takes a lot of toil out of the system right it reads stuff it

synthesizes it summarizes and then it generates it in a way that you want it to generate you can you know make it or

make it shorter make it longer give me a photo you know get it to illustrate I tell you my RMS are saying

to me my Bankers are saying to me wow the pitch book it generates excellent the call reports that it does excellent

the way it's screens for adverse news or or or updated corporate news fabulous so at DBS we've come up with a way to

look at generative AI both vertically and horizontally so vertically we said what

are the vertical use cases right and this is role base so if I'm a banker a relationship manager what can I use

generative AI for pitchbooks client screening ideation right and if I'm a call center

employee what can can I use it for co-pilot when the client calls hey my card is missing immediately it searches

for the client information the transactions and you can answer everything the first call resolution

with this co-pilot went up very sharply and in fact now most of our calls are resolved in the first call so when that

happened I said to my team hey can you even have no calls because if your generative AI is so good then frankly it

should go to the client automatically in their post login to tell them oh I can see you lost your card or I can see your

transaction got stopped how can I help right and um for software developers whether it's GitHub

co-pilot um a lot of the code now testing product acceptance testing uat sit can be done by generative Ai and in

fact people say the quality of the code is pretty darn good and especially if you're trying to move from the main

frame to to micro to to to new stack how many young Engineers know how to do Cobalt very few but the generative AI

can help so many many use cases on the vertical side and then the horizontal what we call mile wide inch deep it's

basically what everybody in the in the organization can use and so we call this some sort of um Enterprise knowledge

base right if you think about it we have 40,000 people people working for us globally some are new some are old some

know this some don't know this everyone's knowledge is different but when you have a safe place for them to

go in and question and they can get exactly the precise answer that you need gosh that saves you a lot of

time and so in terms of building use cases in terms of um getting the right guard rails out this is important

because with the use a generative AI comes a lot of questions around is it dangerous does it

hallucinate is it poisonous data and how would it make us you know give the wrong advice or or was still are we using data

against the customer and here we have come up with something called a pure framework p e

that means whatever data we use it has to to be pure and what is pure pure means means P there's a purpose to

it you it's unsurprising or rather I say unembarrassed to the

customer and R it's respectful to the customer and E it's explainable so with this pure framework

our governance of data stewards we've got the whole team who do this right we'll make sure that the models that we

use the data input is it correct is it pure is it explainable is it embarrassing it's not embarrassing we

make sure we've got these guard rails in place and one of the things that we still do is we have the human in the

loop right so don't worry whilst jobs are going to change humans ain't going to go anywhere humans will still have a

role to play right and and so having that human in the loop is important um the other sort of Forward Thinking

Tech I wanted to talk about is the blockchain I'm just a quick show of hands here I don't know

um we've come through a crypto winter right everybody who you know the the the crypto Market sizing was three trillion

at the peak went down to 1 trillion in the crypto winter and now back up to about 2.8 or so trillion

and crypto might be very volatile but for DBS the real value comes from the blockchain the

DLT the the DLT technology We Believe has a real use case and so crypto winter or not we want to be at the Forefront of

using the blockchain for financial services so we started this actually we started doing this because we saw a lot

of our consumers especially our younger clients buying Bitcoin and ethereum using their credit card and then we

thought oh do they know who they're transacting with do they know whether it's a safe exchange what if they've

been scammed what if you know they lose their money and where do they customize their coins so we thought hey DBS being

one of the world's safest Banks we should offer them this custody service so we started with the custody service

in mind and then when that happened we said oh maybe they want to trade digital tokens with us maybe they want to go on

ramp off-ramp may maybe they want to have Fiat so we said okay we'll build an exchange and then we realized if we're

going to go and do blockchain we better know how to originate which means if in future Financial assets are going to get

tokenized then we better know how to tokenize them right and that means having this origination and insurance

case capabilities so that's where we started we did the whole Suite from origination

and issuance to trading and execution and custody as well as settlements and clearing so we've got the whole

Suite honestly the issuance and origination part has been very slow we did our first Bond we tried to do real

estate it's been very slow because of the lack of market makers but it's there and for trading and execution we've gone

up and down with the market but what has been interesting for us is really the use cases for the DLT technology in

areas of purpose bound money crossb payments and tokenization of Treasury

functions so let me unpeel that so we started a joint venture with JP Morgan tamas stard uh and other Banks

it's called Paro and this is a blockchain based crossb payments platform why is it useful well number

one it's 247 now with platforms everyone's 247 you want to pay cross border and not

wait for banks to open on Mondays right so 24/7 it's Atomic which means it s it's it's Atomic

settlement it's instant and it's prevalidated that means both sides can say this is compliant it's AML

kyc checked we can go ahead and pay go ahead and receive so that's efficient so we see

that as a great use case and the other is tokenized treasury so two weeks ago my colleagues

SE it right here announced that we tied up with ant Financial to help them to tokenize their

treasury tokens in order that they can combine both their treasury tokens for 247 settlements it enables them also to

do liquidity management virtual accounts with all the different merchants and 247 settlements so I see this as a great

use case as the world gets most digitized as platform companies get bigger as more and more trade gets done

digitally this will be a great use case another purpose bound money governments when they give out vouchers

or when they give out Aid packages they want to make sure that the money is being put to good use so Singapore

government for example gives out a lot of help to students you don't want the student to start using that money to buy

beer but you can make the purpose bound money uh be specific for specific use cases you can only buy books or

technology or educated education stuff so I think the blockchain use case is is exciting it's growing and we want

to be at the Forefront and you have to work with with with governments cbdcs as well stable coins now there's more and

more of a growing interest in this whole blockchain ecosystem and now with Regulators on board central banks on

board I think this use case could start to grow okay let me get to the the middle of the

the middle segment which is putting the customer first everyone says this right we learned from all the best tech

companies they're all seeing things from the customer lens but let me tell you for banks we're not a pure digital

business we're a fital business we've got humans we've got branches and we've got digital offerings as well but guess

what the digital branch is the biggest channel right for everyone and for us being an Asian Bank

we wanted to figure out how do we bring an Asian level of service standards to our

customers when I first joined DBS in 2010 we were rated the worst bank for service in Singapore and that's

Singapore's biggest bank that was not good so I remember we were hauled to to a

place in Changi which is near the airport called SQ service quality run by our friends from Singapore Airlines and

they are well known for giving the best service in the world and this is where we learned what it means to give Asian

level of service and Hospitality to our customers whether physically or digitally so we came up with this thing

called not red I'm wearing red red is our DBS corporate colors but red stands for respectful easy to deal with and

dependable that's all you need you need to be red to your customers you need to be respectful easy to deal with

dependable so we don't do anything today without putting the customer in the center that means in our meeting rooms

we've got a chair that's just in red and that chair is the customer Advocate and we don't do any new

products without what we call a 4D framework and what is the 4D framework it's putting the customer in

the discovery discovery mode right what does customer want does he want a mortgage he may not even know what a

mortgage is but he wants to buy a house so start with the Journey of buying a house right that's a customer want um

insurance he may not want insurance but he wants to protect his loved ones start with that journey of protecting his

loved ones so doing the customer Discovery is important in fact

sacran then you Cod and you deliver you de sorry you Cod and you develop and that's the second D and then

you deliver then you define and then you deliver so this is the 4D framework

sorry I should have been disc discover Define develop and deliver in that in that

order and very often what you know we all have unconscious bias don't we we will all say oh uh I know what the

customer wants but very often the data might tell you otherwise let me share a

story when it comes to Foreign Exchange we all thought the client just wants a cheap rate right just give them a good

offer online they'll take it right No in fact we did a lot of AB testing always be guided by the data and we

realized that the data was telling us it wasn't just the price it wasn't the what or the how much or the who it was the

how it was how you communicate to the clients if we just flash hey cheap FX rate we had so so response but if we

flashed hey we missed you come back we haven't seen you in a while here's a great rate to get you back that

worked best so another way that we've changed our body was we've come up with

something called managing through Journeys this was a lot harder to do here we manage through horizontals not

verticals and what does that mean it means having different parts of the bank be it product be it you know the the the

relationship manager be IT Tech be it Ops be it compliance whatever come together in a horizontal organization

and everyone is United by a common goal and so if you manage through Journeys then the journey could be account

opening it could be customer transaction it could be Retail Services it could be documentary trade whatever the journey

is come up with the journey and then we manage through

dashboards we have drivers levers and outcomes and if your driver and your lever and your outcomes are not working

we keep changing but everyone in that horizontal group is driven by the same outcomes we want the customer turnaround

to be lower we want the customer satisfaction to be higher we want more new clients to come on to the product we

want more new clients to come to the bank etc etc so when I go for these mtj meetings with my group

we use dashboards and they use the dashboards to tell me ah this is the outcome good bad what's the lever what's

the driver is it working not working so completely different way of managing completely different way of evaluating

our staff and you know what it's really broken the vertical silos between businesses and Tech and Ops it's made

everyone just focus full-heartedly on customer outcomes and every everyone aligned it's been very difficult but I

think it's gone the long way in helping us reimagine the customer journey and using data and dashboards and control

towers and using Innovation and constantly reiterating experimenting and and getting

better last is the culture so what is culture

culture is a common Behavior you know of your organization it's what binds the teams together it's a common purpose but

creating the right culture is so important and it's harder when you have 38 to 40,000

people and we wanted to create a startup culture in a 38 to 40,000 strong organization how do you do that you know

it's so hard and so we said well you know I I remember my first meeting with the the design team and they said sushan

you're such a hippo and I said what I'm fat what you know and they said no a hippo hipo is a highest paid person in

the room or highest paid person in the office and you're behaving like one because I said hey my ux has to be red

and black that's the DBS colors and they were upset with me because the human Center laboratory was saying the

customers didn't like red and black on their on their screens they like yellow and

green and so they went at Great to Great Lengths to show me what the human Center Design Lab data was telling them and I

said by all means kill red and black go with what the customer wants right and so this is where we came up with

something called the raccoon see the the picture on the right looks like a raccoon but we spelled it with a rack as

in racking ball because we wanted our teams the young people in the room the people who face

the customers every day the young technology guy right the analysts the smart engineer whoever to have a voice

in the room and not be dominated by the hippo or by the boss and so if a young person were to

come up to to me and in the middle of a meeting and say I want to Recon your idea it is the safe space to say go

ahead wreck my idea raccoon me and we found that it was so effective because when everyone said hey I have a raccoon

moment they give we give them a voice and if there are no raccoons we insist come on guys raccoon

us and with every meeting we have a meeting Mojo a mo is the m o meeting owner a Joe is a joyful

Observer normally the Joe is a virtual participant the mo is the person that calls the

meeting and this is where we try to make the meeting more inclusive the joyful Observer could say nah too much

dominance from one person the boss tone down more voices in the room and he will have or he or she will have every Power

to do so and the mo has to speak has to stick to an agenda and a timeline and be inclusive and

encourage descent creating a learning organization we've got a lot of different um ways of

learning different campuses and frankly very unconventional ways of learning when we tried to do when we tried to get

everyone to learn about Ai and machine learning it was during Co you know what we did we did this thing I I can't code

so every one of us we took part in something called The Deep racer AWS deep racer game and we were taught how to cut

and paste code stick it into your car and after the races you go and that's when we we learned that you could use

code and get the machine to learn oh the there's a left turn there there's a right turn there and get your car to go

faster and we encourage failure we encourage Innovation we encourage Rec croud Source ideas we've got failure

Awards Daredevil Awards Etc and um it's it's this culture of innovation and and groundup Innovation

that really you know puts us ahead finally it's what we call our pride culture actually this is a common thing

this is something that hasn't changed this we've had for years and being a Purpose Driven organization Pride right

purpose relationship that Innovative decisive and E everything fun we couldn't find something that began with

e maybe enjoyment but not many organizations put fun as something in that culture but we do because you want

people to have fun you want you we spend so much time in the in in in in at work we want them to be able to create

memories that last a lifetime so this is important so I will end with a slide that I love which is a quote from Peter

draer which is culture eats strategy for breakfast you can have the best strategy and the best team but if you don't have

the right culture you're not going to make it so at the end of the day create the right culture create the right us

and vision for your team and you'll do really well so with that thank you very much

[Applause]