3024: Unleashing the Power of AI in Banking: DBS Bank's Strategy for Success

[00:00:03] [SPEAKER_01]: How is generative AI poised to transform the financial sector?

[00:00:09] [SPEAKER_01]: And what ethical considerations must be addressed to harness its full potential?

[00:00:15] [SPEAKER_01]: Well today I'm joined by the Chief Data and Transformation Officer at DBS Bank.

[00:00:21] [SPEAKER_01]: And together we're going to explore the groundbreaking impact of generative AI on

[00:00:26] [SPEAKER_01]: banking.

[00:00:27] [SPEAKER_01]: And we'll also learn more about how DBS Bank is embracing AI to enhance productivity,

[00:00:32] [SPEAKER_01]: reduce administrative burdens and even improve customer experiences.

[00:00:39] [SPEAKER_01]: But most importantly, all while maintaining rigorous governance and ethical standards.

[00:00:46] [SPEAKER_01]: And what does that mean for the future of banking and the roles of those within it?

[00:00:51] [SPEAKER_01]: These are just a few other things that we're going to explore today.

[00:00:56] [SPEAKER_01]: Delivering daily content to 140,000 of you wonderful monthly listeners across the

[00:01:02] [SPEAKER_01]: I don't want to take all the credit here because it wouldn't be possible without the backing

[00:01:06] [SPEAKER_01]: of our dedicated sponsors and partners.

[00:01:08] [SPEAKER_01]: And I also want to shine a light on the fact that legacy managed file transfer tools

[00:01:12] [SPEAKER_01]: are looking dated, they often lack the security that today's remote workforce demands.

[00:01:17] [SPEAKER_01]: And companies that continue relying on that outdated tech, they're in danger of

[00:01:20] [SPEAKER_01]: putting their sensitive data at risk.

[00:01:23] [SPEAKER_01]: Attention IT professionals, are you tired of juggling these multiple servers for

[00:01:27] [SPEAKER_01]: secure file sharing, integrated shared folders and email plus a comprehensive REST API?

[00:01:33] [SPEAKER_01]: Kiteworks simplifies your workflow.

[00:01:35] [SPEAKER_01]: For administrators, you can experience unmatched functionality and integration that traditional

[00:01:41] [SPEAKER_01]: MFT servers just can't touch.

[00:01:43] [SPEAKER_01]: Step into the future of secure managed file transfer with Kiteworks by going to Kiteworks.com

[00:01:49] [SPEAKER_01]: to get started.

[00:01:50] [SPEAKER_01]: That's Kiteworks.com.

[00:01:51] [SPEAKER_01]: And remember, Kiteworks is also FedRAMP, moderate authorised.

[00:01:55] [SPEAKER_01]: Thank you for your patience today.

[00:01:56] [SPEAKER_01]: This is the moment you've been waiting for.

[00:01:58] [SPEAKER_01]: It's time to welcome my guest onto the show.

[00:02:02] [SPEAKER_01]: So a massive warm welcome to the show.

[00:02:04] [SPEAKER_01]: Can you tell everyone listening a little bit who you are and what you do?

[00:02:08] [SPEAKER_00]: So my name is Dhritri Prachmatyar.

[00:02:10] [SPEAKER_00]: I am the Chief Data and Transformation Officer of DBS Bank in Singapore for the

[00:02:15] [SPEAKER_00]: entire group.

[00:02:17] [SPEAKER_00]: My role entails largely looking around corners, figuring out what's coming and then

[00:02:23] [SPEAKER_00]: sure that we understand it really well, bringing it into the organisation, understanding the

[00:02:28] [SPEAKER_00]: implications and then determining what books of work we need to develop and then therefore

[00:02:34] [SPEAKER_00]: deploy into the organisation so that we're future ready at every given point or on

[00:02:39] [SPEAKER_00]: an ongoing basis.

[00:02:41] [SPEAKER_00]: What I cover, I cover the entire data agenda for the bank.

[00:02:46] [SPEAKER_00]: And so data is a tool for transforming the way we work and the way work gets

[00:02:51] [SPEAKER_00]: done. So that be it through AI ML, be it through generative AI, data governance,

[00:02:55] [SPEAKER_00]: data management.

[00:02:56] [SPEAKER_00]: I'm also responsible for what is in the outside world known as Agile at scale, but

[00:03:02] [SPEAKER_00]: internally we call it managing through journeys.

[00:03:04] [SPEAKER_00]: This is where we move the traditional silos that exist in banks and have existed

[00:03:08] [SPEAKER_00]: for hundreds of years into a more horizontal structure so that it addresses actually

[00:03:14] [SPEAKER_00]: customer journeys rather than organisation efficiency.

[00:03:18] [SPEAKER_00]: And that does not mean we're adding in inefficiency at all.

[00:03:21] [SPEAKER_00]: In fact, what we've seen is a lot it gets very efficient when we arrange ourselves

[00:03:25] [SPEAKER_00]: around the jobs for the customer.

[00:03:28] [SPEAKER_00]: Quality improves, stickiness improves and therefore loyalty improves and revenues

[00:03:32] [SPEAKER_00]: improve.

[00:03:34] [SPEAKER_00]: That's another part of what I do.

[00:03:36] [SPEAKER_00]: I'm also responsible for customer experience and employee experience for banks,

[00:03:40] [SPEAKER_00]: the future of work.

[00:03:41] [SPEAKER_00]: And finally, I'm also responsible for the innovation agenda of the bank.

[00:03:45] [SPEAKER_00]: So how do we accelerate and how do we influence innovation?

[00:03:51] [SPEAKER_00]: I have a team of people that work with the entire organisation to help them

[00:03:56] [SPEAKER_00]: innovate within their own spaces.

[00:03:59] [SPEAKER_00]: And also innovation is not supposed to be done by four or five people.

[00:04:02] [SPEAKER_00]: I mean, to be an innovative company, everybody needs to be focused on innovating

[00:04:06] [SPEAKER_00]: services and products.

[00:04:07] [SPEAKER_00]: So that's largely my remit if you like.

[00:04:10] [SPEAKER_01]: Love that. And there are so many different reasons I was excited to get you on

[00:04:12] [SPEAKER_01]: the podcast today because there is a lot of hype around AI at the moment and

[00:04:17] [SPEAKER_01]: different industries have different levels of, let's say, cautiousness around

[00:04:21] [SPEAKER_01]: adopting this technology.

[00:04:22] [SPEAKER_01]: And I've spoken to many banks on this podcast and even had one last year or

[00:04:27] [SPEAKER_01]: it was earlier this year, a tech conference.

[00:04:29] [SPEAKER_01]: They were scheduled to meet me at the conference, have a podcast, but

[00:04:32] [SPEAKER_01]: their senior management called time on the podcast before we even recorded

[00:04:35] [SPEAKER_01]: because they weren't comfortable with having a conversation about AI.

[00:04:39] [SPEAKER_01]: So I see a lot of different attitudes towards it.

[00:04:43] [SPEAKER_01]: So I'm curious, is someone right in the heart of this space?

[00:04:45] [SPEAKER_01]: Can you elaborate on how you see the transformational potential of

[00:04:49] [SPEAKER_01]: generative AI and the financial sector, particularly in terms of things

[00:04:54] [SPEAKER_01]: like cost efficiency and savings?

[00:04:56] [SPEAKER_01]: Because it's not just about look at the shiny new technology.

[00:04:59] [SPEAKER_01]: We're solving real problems with it, right?

[00:05:02] [SPEAKER_00]: So maybe you talked about people being scared to talk about it.

[00:05:05] [SPEAKER_00]: You talk about the potential.

[00:05:07] [SPEAKER_00]: So in terms of, I'd say first and foremost, one shouldn't be scared of this.

[00:05:11] [SPEAKER_00]: Right?

[00:05:11] [SPEAKER_00]: It's here.

[00:05:12] [SPEAKER_00]: It's not that it's suddenly going to disappear and we've got to embrace it

[00:05:17] [SPEAKER_00]: and embrace it responsibly and make the best of what, you know,

[00:05:21] [SPEAKER_00]: the advancement of this technology.

[00:05:22] [SPEAKER_00]: And I think there's a lot of benefits.

[00:05:24] [SPEAKER_00]: Now we've got to be careful.

[00:05:25] [SPEAKER_00]: There are downsides as well.

[00:05:26] [SPEAKER_00]: These we need to be very conscious about and manage them correctly.

[00:05:29] [SPEAKER_00]: Right now in terms of potential, I think it's huge.

[00:05:33] [SPEAKER_00]: It will really change the way work gets done.

[00:05:36] [SPEAKER_00]: And change managed correctly, it has huge potential.

[00:05:40] [SPEAKER_00]: The ability to take knowledge and unstructured data from many different

[00:05:44] [SPEAKER_00]: places and bring it in and make it searchable, make it understandable,

[00:05:50] [SPEAKER_00]: models synthesize the stuff, summarize it.

[00:05:52] [SPEAKER_00]: I think that will remove so much toil from the day-to-day work of people

[00:05:57] [SPEAKER_00]: that it will free up significant capacities for people to be able to do

[00:06:01] [SPEAKER_00]: higher value stuff.

[00:06:03] [SPEAKER_00]: Right?

[00:06:03] [SPEAKER_00]: So, I mean, I don't know about your role, but in my role and in a lot of

[00:06:08] [SPEAKER_00]: my colleagues, we spend a lot of time looking for stuff.

[00:06:11] [SPEAKER_00]: And a large organization is full of very rich knowledge that really

[00:06:18] [SPEAKER_00]: helps people do stuff now.

[00:06:19] [SPEAKER_00]: But to get to it is quite difficult.

[00:06:22] [SPEAKER_00]: There's a lot of formats to start with, different repositories where they sit.

[00:06:28] [SPEAKER_00]: Some are not really holistic documents or holistic information.

[00:06:33] [SPEAKER_00]: It's one dimension or one prism.

[00:06:36] [SPEAKER_00]: Bringing all this together and being able to synthesize what's there and

[00:06:41] [SPEAKER_00]: make sense of it and be able to query from different angles, I think will

[00:06:45] [SPEAKER_00]: reduce a significant amount of toil, significant amount of process steps that

[00:06:49] [SPEAKER_00]: exist in organizations and therefore allow for these efficiencies to be

[00:06:54] [SPEAKER_00]: invested in the right place.

[00:06:56] [SPEAKER_00]: And for me, the right places are how can we up our customer experience?

[00:07:00] [SPEAKER_00]: How can we up our employee experience?

[00:07:02] [SPEAKER_00]: And then as a result of that, how can we have better outcomes financial

[00:07:05] [SPEAKER_00]: or otherwise?

[00:07:06] [SPEAKER_01]: And as you said a moment ago there, there were so many great

[00:07:09] [SPEAKER_01]: opportunities, but equally there are other areas that need to be

[00:07:12] [SPEAKER_01]: managed and you need to balance that power and responsibility accordingly.

[00:07:16] [SPEAKER_01]: So with the advancements in generative AI, how do you see banks

[00:07:21] [SPEAKER_01]: balancing that power of this technology with the responsibility it entails,

[00:07:25] [SPEAKER_01]: especially in terms of things like ethical considerations, because a big

[00:07:29] [SPEAKER_01]: talking point across so many different sectors right now.

[00:07:32] [SPEAKER_00]: I know that there are certain regulators that are taking it very

[00:07:35] [SPEAKER_00]: seriously.

[00:07:36] [SPEAKER_00]: In Singapore, the MS is working on FIT and Veritas, which are

[00:07:41] [SPEAKER_00]: principles that govern the use of AI and ML and use of data.

[00:07:46] [SPEAKER_00]: So we know that the regulators are engaged in this, but it's an

[00:07:48] [SPEAKER_00]: evolving space.

[00:07:50] [SPEAKER_00]: What are, how are we thinking about it?

[00:07:51] [SPEAKER_00]: So we think about it along the lines of fairness, ethical accountability

[00:07:56] [SPEAKER_00]: and transparency.

[00:07:57] [SPEAKER_00]: And we've got our own standards because this isn't the AI MS journey

[00:08:02] [SPEAKER_00]: is not new for us.

[00:08:03] [SPEAKER_00]: We started way back in 2018, 2019 when what is today known as

[00:08:08] [SPEAKER_00]: traditional AI was really the cutting edge.

[00:08:11] [SPEAKER_00]: So we did develop frameworks around how to be responsible when using

[00:08:17] [SPEAKER_00]: data and when using AI.

[00:08:20] [SPEAKER_00]: And that gave us a good basis to then extend it to generative AI.

[00:08:25] [SPEAKER_00]: Right.

[00:08:26] [SPEAKER_00]: And we are very deliberate.

[00:08:28] [SPEAKER_00]: One, we focused largely on co-pilots because we're not yet quite

[00:08:32] [SPEAKER_00]: comfortable around autonomous AI.

[00:08:36] [SPEAKER_00]: We don't feel it's ready to be left by itself to customers or even

[00:08:41] [SPEAKER_00]: to employees at the moment.

[00:08:42] [SPEAKER_00]: Right.

[00:08:42] [SPEAKER_00]: Hallucination, appropriateness, bias, all these things still have to be

[00:08:47] [SPEAKER_00]: resolved.

[00:08:48] [SPEAKER_00]: So, as I say, our focus has largely been around co-pilot and we're very

[00:08:52] [SPEAKER_00]: deliberate about every single use case.

[00:08:55] [SPEAKER_00]: And so we have several levels of validation that the solutions that

[00:09:00] [SPEAKER_00]: we're putting out to out there are fit for purpose from a risk

[00:09:05] [SPEAKER_00]: perspective.

[00:09:06] [SPEAKER_00]: And it goes right to the top where they are senior people.

[00:09:09] [SPEAKER_00]: Again, the top post committee, which we call responsible data use is

[00:09:13] [SPEAKER_00]: chaired by me and I have my peers from compliance, legal, HR, and various

[00:09:19] [SPEAKER_00]: other units where we will deliberate whether the use case has covered

[00:09:25] [SPEAKER_00]: every single side of every single angle of risk.

[00:09:29] [SPEAKER_00]: And being that deliberate has allowed us to maybe not be so fast at

[00:09:33] [SPEAKER_00]: going out with these solutions, but whatever we have out there,

[00:09:36] [SPEAKER_00]: we've got over 20 use cases in production now.

[00:09:40] [SPEAKER_00]: They're well thought through.

[00:09:41] [SPEAKER_00]: They're being used and we're seeing that we're not seeing anything that

[00:09:45] [SPEAKER_00]: worries us in terms of the gen AI going its own way.

[00:09:48] [SPEAKER_00]: Right.

[00:09:49] [SPEAKER_00]: So we are very deliberate.

[00:09:51] [SPEAKER_00]: I know some organizations that have actually gone out and put this out

[00:09:55] [SPEAKER_00]: in front of customers, but they're not really regulated industries.

[00:10:01] [SPEAKER_00]: So banking, healthcare is a heavily regulated and you're not allowed to

[00:10:05] [SPEAKER_00]: be wrong any time when you regulate and we'll come after you.

[00:10:08] [SPEAKER_00]: So we've got to be very deliberate about it.

[00:10:09] [SPEAKER_00]: And until we're absolutely comfortable that this technology allows us to be

[00:10:14] [SPEAKER_00]: at the same standard, we will be very careful before we put it out there.

[00:10:19] [SPEAKER_00]: But as far as co-pilot is concerned, I think we've made a lot of

[00:10:22] [SPEAKER_00]: progress and we continue to accelerate the use cases across the organization.

[00:10:28] [SPEAKER_01]: And there's a lot of misconceptions out there at the moment that AI is

[00:10:31] [SPEAKER_01]: going to replace things and people go all in on AI or all in on humans.

[00:10:35] [SPEAKER_01]: And for me, I think the real magic happens when there's this seamless

[00:10:38] [SPEAKER_01]: synergy between data, humans and machines all working together in ways

[00:10:43] [SPEAKER_01]: that they can't do completely on their own in their own rights.

[00:10:46] [SPEAKER_01]: So what steps do you think are crucial to unlock the full potential

[00:10:50] [SPEAKER_01]: of this technology in this way?

[00:10:52] [SPEAKER_00]: So first of all, is it going to have change?

[00:10:55] [SPEAKER_00]: Is it going to impact roles?

[00:10:57] [SPEAKER_00]: Yes, but it will impact roles.

[00:10:59] [SPEAKER_00]: It won't necessarily impact people.

[00:11:01] [SPEAKER_00]: Right.

[00:11:01] [SPEAKER_00]: I think as this evolves and you get adoption and we find the best

[00:11:06] [SPEAKER_00]: use this for us and also find areas where we don't use it.

[00:11:10] [SPEAKER_00]: This will result into maybe some activities and some certain types of

[00:11:15] [SPEAKER_00]: roles being impacted, but it's also going to create a lot of new roles.

[00:11:20] [SPEAKER_00]: I mean, I was talking with somebody earlier today, managing the

[00:11:24] [SPEAKER_00]: outputs of this generative capability.

[00:11:26] [SPEAKER_00]: That's a new job that's going to have to be created.

[00:11:29] [SPEAKER_00]: And you're going to need supervisors for these models.

[00:11:32] [SPEAKER_00]: Right.

[00:11:32] [SPEAKER_00]: And those supervisors are going to have to be humans.

[00:11:34] [SPEAKER_00]: And so that creates a new role altogether.

[00:11:38] [SPEAKER_00]: And that brings it very nicely to your second point or your second question,

[00:11:42] [SPEAKER_00]: which is this interaction.

[00:11:44] [SPEAKER_00]: Yes.

[00:11:44] [SPEAKER_00]: That's where the magic is.

[00:11:46] [SPEAKER_00]: Right.

[00:11:46] [SPEAKER_00]: Again, if you're use case-driven and you are, you're absolutely

[00:11:50] [SPEAKER_00]: clear on what outcomes you want and you understand the risks and

[00:11:53] [SPEAKER_00]: mitigation of those risks, that balance is important to get this.

[00:11:57] [SPEAKER_00]: The beauty, as you say, of these three coming together.

[00:12:01] [SPEAKER_00]: It's not just necessarily generative AI that is going to bring magic

[00:12:04] [SPEAKER_00]: actually sink the interaction of the generative, which is unstructured data.

[00:12:09] [SPEAKER_00]: And structured data, which is, which is now referred to as traditional AI

[00:12:13] [SPEAKER_00]: or traditional AI ML, these coming together and the abilities you've

[00:12:18] [SPEAKER_00]: got in terms of compute and people being able to instruct how this works.

[00:12:24] [SPEAKER_00]: I think that's going to be where real magic will happen.

[00:12:28] [SPEAKER_01]: 100% with you there.

[00:12:30] [SPEAKER_01]: And another role that I keep hearing about are data stewards.

[00:12:33] [SPEAKER_01]: So I'm curious, how do you see data stewards playing a role in ensuring

[00:12:37] [SPEAKER_01]: that accuracy, quality and security of AI models?

[00:12:41] [SPEAKER_01]: And do you think this role is going to become increasingly important?

[00:12:44] [SPEAKER_01]: It certainly feels that way.

[00:12:46] [SPEAKER_00]: At least in our organisation, it's a very important role and

[00:12:48] [SPEAKER_00]: has been for several years.

[00:12:50] [SPEAKER_00]: So data steward in the traditional sense was a much easier job

[00:12:54] [SPEAKER_00]: in traditional AI ML, right?

[00:12:57] [SPEAKER_00]: You've got to make sure you've got right level of metadata.

[00:12:59] [SPEAKER_00]: You're going to make sure you've got right link lineages.

[00:13:02] [SPEAKER_00]: You've got your processes in place.

[00:13:03] [SPEAKER_00]: You've got a platform in place.

[00:13:04] [SPEAKER_00]: You've got your tooling in place and they ensure that the right kind

[00:13:06] [SPEAKER_00]: of governance is there, who gets access to data, who doesn't,

[00:13:10] [SPEAKER_00]: and for what type of use cases, right?

[00:13:11] [SPEAKER_00]: It's a traditional data steward job.

[00:13:14] [SPEAKER_00]: Now that was fine because it's data, structured data is actually

[00:13:21] [SPEAKER_00]: quite rich in its way and it can be misused if you need data

[00:13:26] [SPEAKER_00]: for your job, you should get it.

[00:13:28] [SPEAKER_00]: Right.

[00:13:28] [SPEAKER_00]: But everybody having all sorts of data was, is it from a risk perspective,

[00:13:34] [SPEAKER_00]: it needed to be mitigated and where then the data steward role became stronger.

[00:13:38] [SPEAKER_00]: Now when you bring in generative AI, that changes the game a little

[00:13:42] [SPEAKER_00]: because now it isn't data, it's knowledge.

[00:13:47] [SPEAKER_00]: And in advanced companies that are very advanced in the use of data,

[00:13:50] [SPEAKER_00]: typically you would have a data pool or a data lake and then you got all

[00:13:54] [SPEAKER_00]: your tooling and your governance and your processes and all that off there

[00:13:57] [SPEAKER_00]: and people work off a common platform.

[00:14:00] [SPEAKER_00]: The problem with knowledge in an institution, it's everywhere.

[00:14:03] [SPEAKER_00]: It's in a Word document and that Word document is in your

[00:14:06] [SPEAKER_00]: shared folder, right?

[00:14:08] [SPEAKER_00]: Now how do you bring all of that, right?

[00:14:11] [SPEAKER_00]: And allow generative AI to be able to make sense of it, synthesize

[00:14:14] [SPEAKER_00]: it and make it searchable is, it's going to be a different role from

[00:14:19] [SPEAKER_00]: the structured AI ML data steward.

[00:14:22] [SPEAKER_00]: So I see the data steward role evolving as we go forward.

[00:14:25] [SPEAKER_00]: It might not even be the same data steward for AI ML and for unstructured

[00:14:30] [SPEAKER_00]: data, it may well be, but let's see how that goes.

[00:14:33] [SPEAKER_00]: But I think that that role is going to be, will evolve and it

[00:14:35] [SPEAKER_00]: becomes extremely important.

[00:14:36] [SPEAKER_00]: Now it's a question of not just the data, it's now about knowledge.

[00:14:40] [SPEAKER_00]: I mean, and information that is completely unstructured and

[00:14:44] [SPEAKER_00]: everybody's got knowledge, right?

[00:14:46] [SPEAKER_00]: I'm working on a document, for example, for this podcast that's

[00:14:50] [SPEAKER_00]: sitting on my OneDrive and there's important information there.

[00:14:54] [SPEAKER_00]: If somebody else, maybe Onkel who works in my unit or in marketing

[00:14:58] [SPEAKER_00]: or something would like to know what I say, then my documents got

[00:15:02] [SPEAKER_00]: to be discoverable, it's got to be searchable.

[00:15:06] [SPEAKER_00]: So how this will work is stuff that needs to be figured out.

[00:15:13] [SPEAKER_00]: And so the data steward will have to determine, for example,

[00:15:17] [SPEAKER_00]: what kind of metadata do you want on each document, for example?

[00:15:22] [SPEAKER_00]: Right?

[00:15:23] [SPEAKER_00]: While you can describe metadata very strictly for structured data,

[00:15:28] [SPEAKER_00]: right?

[00:15:29] [SPEAKER_00]: And make it discoverable.

[00:15:30] [SPEAKER_00]: It's very difficult because if you've got a credit policy document as a

[00:15:35] [SPEAKER_00]: salesperson, my view of the document is very different from the credit

[00:15:39] [SPEAKER_00]: person than from the compliance person.

[00:15:41] [SPEAKER_00]: So there are some challenges that need to be figured out, but I think the

[00:15:45] [SPEAKER_00]: data steward role will evolve and it will become a little bit more

[00:15:50] [SPEAKER_00]: complex than it is today, not to say that it's not complex as it is.

[00:15:54] [SPEAKER_01]: And earlier in our conversation, you were talking about how heavily

[00:15:58] [SPEAKER_01]: regulated the financial and banking industry is and as pressures

[00:16:02] [SPEAKER_01]: mount for responsible and ethical AI use.

[00:16:06] [SPEAKER_01]: I'm curious, what kind of measures do you put into place to ensure

[00:16:10] [SPEAKER_01]: that these standards are met?

[00:16:11] [SPEAKER_01]: Because as you said a few moments ago, there is no failure.

[00:16:14] [SPEAKER_01]: There's no option to get it wrong once it's got to be right from

[00:16:19] [SPEAKER_01]: the outset.

[00:16:19] [SPEAKER_01]: So what kind of measures do you have in place here?

[00:16:21] [SPEAKER_00]: So at the moment we are focusing on everything to be copilot and there

[00:16:25] [SPEAKER_00]: must be a human in the loop.

[00:16:27] [SPEAKER_00]: That's at the moment, the only way we think it's going to work.

[00:16:30] [SPEAKER_00]: And it becomes then the responsibility of the human to ensure that

[00:16:34] [SPEAKER_00]: what's come out of it is correct.

[00:16:36] [SPEAKER_00]: Now at this level of maturity where GenAI is still picking up, right?

[00:16:43] [SPEAKER_00]: The human knows a lot more, right?

[00:16:45] [SPEAKER_00]: And so therefore can become a checker if you like, but as more

[00:16:48] [SPEAKER_00]: and more processes, steps, work gets, you use GenAI for it.

[00:16:55] [SPEAKER_00]: We're going to have to think about how we are smarter about it.

[00:16:58] [SPEAKER_00]: But the use cases are quite specific and the subject matter experts are humans.

[00:17:05] [SPEAKER_00]: And so they will continue to have a look at any output.

[00:17:08] [SPEAKER_00]: And then we go, for example, if I'm a sales guy advising you on some

[00:17:12] [SPEAKER_00]: product, I will get the Genitive AI to understand your profile and read

[00:17:16] [SPEAKER_00]: through several CIO reports and analyst reports that are on

[00:17:21] [SPEAKER_00]: Cassandra Baker recommendation.

[00:17:22] [SPEAKER_00]: And I will read it and I will say, okay, you know what?

[00:17:25] [SPEAKER_00]: This type of insurance is right for me.

[00:17:28] [SPEAKER_00]: And it makes sense.

[00:17:29] [SPEAKER_00]: Now, if we left it to the Genitive AI, I think you have the risk of

[00:17:33] [SPEAKER_00]: the wrong product being proposed.

[00:17:35] [SPEAKER_00]: And that becomes a regulatory issue as well, if you were to

[00:17:38] [SPEAKER_00]: complain, the human in the loop is really important in this case, right?

[00:17:41] [SPEAKER_00]: And we'll remain until we're able to figure out that every time we get an

[00:17:47] [SPEAKER_00]: output from the generative capability, that it is a hundred percent.

[00:17:52] [SPEAKER_00]: Right.

[00:17:52] [SPEAKER_00]: And it's not just a hundred percent accuracy because it's not about

[00:17:54] [SPEAKER_00]: it being accurate all the time.

[00:17:56] [SPEAKER_00]: Right.

[00:17:56] [SPEAKER_00]: There are other issues that come about.

[00:17:59] [SPEAKER_00]: One is it's not fully accurate.

[00:18:01] [SPEAKER_00]: So it's partially accurate.

[00:18:03] [SPEAKER_00]: That doesn't mean the rest of it is wrong.

[00:18:04] [SPEAKER_00]: It's just omitting some stuff.

[00:18:06] [SPEAKER_00]: Right?

[00:18:06] [SPEAKER_00]: So that's one thing still that needs to be dealt with.

[00:18:08] [SPEAKER_00]: The other one is it could be factually a hundred percent correct.

[00:18:11] [SPEAKER_00]: But is it appropriate?

[00:18:13] [SPEAKER_00]: Right.

[00:18:13] [SPEAKER_00]: In terms of language and because the LLMs are still a black box, we don't

[00:18:17] [SPEAKER_00]: understand what's going on at the back there.

[00:18:19] [SPEAKER_00]: And it's difficult to explain why it came up with the reason.

[00:18:25] [SPEAKER_00]: Now you can do a lot of engineering through chain of thought and all that

[00:18:28] [SPEAKER_00]: to figure out why it saved what it saved.

[00:18:31] [SPEAKER_00]: Right.

[00:18:31] [SPEAKER_00]: In terms of the facts, but in terms of the language, right.

[00:18:36] [SPEAKER_00]: That you can't control.

[00:18:39] [SPEAKER_00]: That is however we, whichever LLM you use, that's the black box.

[00:18:45] [SPEAKER_00]: Right.

[00:18:46] [SPEAKER_00]: And so we've got to be a little bit careful about.

[00:18:48] [SPEAKER_00]: No, a lot of people focus on hallucination, but it's not, that's

[00:18:50] [SPEAKER_00]: not the only problem that is there.

[00:18:52] [SPEAKER_00]: I think appropriateness bias, any, I think toxicity.

[00:18:57] [SPEAKER_00]: These are things we really need to worry about despite the fact

[00:18:59] [SPEAKER_00]: that the answer might be absolutely a hundred percent factually correct.

[00:19:04] [SPEAKER_00]: Right.

[00:19:04] [SPEAKER_00]: So there's a lot of things outside of hallucination accuracy

[00:19:07] [SPEAKER_00]: that we need to worry about before, before we let this.

[00:19:11] [SPEAKER_01]: So many great points in there.

[00:19:13] [SPEAKER_01]: And for anybody listening that may be sat on the fence at the moment,

[00:19:16] [SPEAKER_01]: may they were a business leader in the industry wanting to learn more

[00:19:19] [SPEAKER_01]: and maybe follow in your footsteps, but want to know more about what

[00:19:23] [SPEAKER_01]: the big problems we solve and how we're making things better here.

[00:19:26] [SPEAKER_01]: Is there anything you can share around how you're leveraging

[00:19:28] [SPEAKER_01]: gen AI to achieve things like productivity gains across systems,

[00:19:33] [SPEAKER_01]: customers and employees, because there seems to be a lot of rich

[00:19:37] [SPEAKER_01]: things that.

[00:19:38] [SPEAKER_00]: Quick example would be our customer service officers in the call centers.

[00:19:43] [SPEAKER_00]: So we've given them capability to be able to transcribe the call,

[00:19:47] [SPEAKER_00]: summarize it and call out in terms of what actions need to be taken

[00:19:52] [SPEAKER_00]: and then populate that into the system so that they can be, maybe

[00:19:55] [SPEAKER_00]: security card renewal or whatever the case may be.

[00:19:57] [SPEAKER_00]: And then it triggers it off straight through also in terms of,

[00:20:01] [SPEAKER_00]: you know, through the transcribing, what's the call,

[00:20:04] [SPEAKER_00]: either it'll make recommendations.

[00:20:05] [SPEAKER_00]: This has resulted into 20% improvement in the dog type, in the handling type

[00:20:11] [SPEAKER_00]: of the customer, which allows then the customer service officer

[00:20:17] [SPEAKER_00]: to be more empathetic, right?

[00:20:19] [SPEAKER_00]: So now they're not rushing to meet their KPI in terms of, I got

[00:20:22] [SPEAKER_00]: another call waiting for me, which they do, but it does give them

[00:20:27] [SPEAKER_00]: the opportunity to recommend products, understand better the customer's

[00:20:31] [SPEAKER_00]: product empathy.

[00:20:32] [SPEAKER_00]: A lot of banking call centers get accused of not having any empathy, right?

[00:20:38] [SPEAKER_00]: And then these are all winners, right?

[00:20:39] [SPEAKER_00]: I mean, if I make you, if you're happy with the way you've been

[00:20:42] [SPEAKER_00]: handled, despite the fact that it was a problem, the likelihood of you

[00:20:45] [SPEAKER_00]: coming back to me is much higher than before.

[00:20:48] [SPEAKER_00]: So we are seeing productivity gains there in our branches.

[00:20:52] [SPEAKER_00]: We're seeing that the branch officers looking for information to serve

[00:20:57] [SPEAKER_00]: customer and type of documentation, product or compliance related.

[00:21:01] [SPEAKER_00]: This is now available at their fingertips with a bot.

[00:21:06] [SPEAKER_00]: And this has reduced, you know, their daily time by nearly 8%, 6 to 8%.

[00:21:12] [SPEAKER_00]: So we're starting to see that again, frees up the time for them

[00:21:15] [SPEAKER_00]: to do quality work with the customer or at the end of day for

[00:21:20] [SPEAKER_00]: their back on fist work and all that.

[00:21:21] [SPEAKER_00]: So we're seeing some very good traction.

[00:21:25] [SPEAKER_00]: And this is what we're after.

[00:21:27] [SPEAKER_00]: Again, as I say, Co-Pilot allows us to make sure that whatever's

[00:21:31] [SPEAKER_00]: coming out is appropriate and correct.

[00:21:33] [SPEAKER_00]: And then we can pass it on to the customer.

[00:21:35] [SPEAKER_00]: So then there's many more.

[00:21:36] [SPEAKER_00]: I mean, working with relationship managers, we're working with sales.

[00:21:39] [SPEAKER_00]: I will give you one more example since you are an example.

[00:21:43] [SPEAKER_00]: For many of the large Co-Pilot relationship managers, one of the

[00:21:47] [SPEAKER_00]: most difficult thing was establishing the ESG standards of an organization.

[00:21:53] [SPEAKER_00]: Right?

[00:21:53] [SPEAKER_00]: And this is typically hidden in about 120 page report that every

[00:21:59] [SPEAKER_00]: company will publish every year.

[00:22:01] [SPEAKER_00]: And we're looking for between 15 and 17 criteria to establish whether

[00:22:05] [SPEAKER_00]: indeed we will, we're willing to do business with these people and do

[00:22:08] [SPEAKER_00]: their ESG, does their ESG philosophy match with our ESG philosophy?

[00:22:14] [SPEAKER_00]: Now, typically you can imagine reading a 120 page document, especially

[00:22:17] [SPEAKER_00]: when it's an ESG report, that's better than sleeping pills, right?

[00:22:21] [SPEAKER_00]: I mean, you will, it's hard.

[00:22:23] [SPEAKER_00]: So it takes, it takes very long.

[00:22:25] [SPEAKER_00]: So we've been able to build capability that will take the entire

[00:22:29] [SPEAKER_00]: ESG report and within a minute, pump out all the criteria that we want,

[00:22:33] [SPEAKER_00]: depending on what industry they are in, thereby significantly reducing

[00:22:37] [SPEAKER_00]: the amount of foil that a relationship manager needs to go through.

[00:22:42] [SPEAKER_01]: And in the face of the age old paranoia of machines versus humans,

[00:22:47] [SPEAKER_01]: it's as old as time itself.

[00:22:49] [SPEAKER_01]: We've seen it for hundreds, if not thousands of years, but

[00:22:51] [SPEAKER_01]: we're still here now.

[00:22:52] [SPEAKER_01]: This time it's AI.

[00:22:54] [SPEAKER_01]: So how is, how are you at DBS Bank addressing some of those concerns

[00:22:59] [SPEAKER_01]: and fostering trust in AI technologies, whether it be your customers or

[00:23:03] [SPEAKER_01]: employees, I'm sure you get this a lot, but how are you dealing

[00:23:07] [SPEAKER_01]: with that initial myth or preconception that some may have?

[00:23:11] [SPEAKER_00]: So if you think, I think there's three levels of impact here, right?

[00:23:14] [SPEAKER_00]: One is where there's augmentation, right?

[00:23:18] [SPEAKER_00]: So I'm going to give you a tool.

[00:23:20] [SPEAKER_00]: It works together with you.

[00:23:21] [SPEAKER_00]: Right.

[00:23:22] [SPEAKER_00]: And if you need to write a better email, just go in, put in your bullet points,

[00:23:25] [SPEAKER_00]: outcomes, very nice, large, you know, saves you 15, 20 minutes

[00:23:28] [SPEAKER_00]: or the kind of stuff.

[00:23:29] [SPEAKER_00]: This is very easy to do.

[00:23:31] [SPEAKER_00]: Right.

[00:23:31] [SPEAKER_00]: And when you tell people and explain to people the capabilities,

[00:23:34] [SPEAKER_00]: people don't ask questions as to, oh, is this going to take my job?

[00:23:37] [SPEAKER_00]: Okay.

[00:23:39] [SPEAKER_00]: Call pilots, right?

[00:23:41] [SPEAKER_00]: And we want to call pilot everything.

[00:23:43] [SPEAKER_00]: And we really do believe actually co-pilot is more internally,

[00:23:46] [SPEAKER_00]: we call it cool worker, right?

[00:23:48] [SPEAKER_00]: So I've got somebody who's I'm going to work with.

[00:23:52] [SPEAKER_00]: Right?

[00:23:53] [SPEAKER_00]: So this has a degree of impact, right?

[00:23:55] [SPEAKER_00]: So the way it really compliments, for example, in the customer service

[00:23:59] [SPEAKER_00]: officers, they love it.

[00:24:00] [SPEAKER_00]: I mean, it was unbelievable.

[00:24:02] [SPEAKER_00]: There were those skepticism at the beginning, but you know, once they

[00:24:04] [SPEAKER_00]: figured it out, they were like, oh, wow, this is great.

[00:24:07] [SPEAKER_00]: Please embed it in our workflow.

[00:24:08] [SPEAKER_00]: And we did and we don't have any complaints.

[00:24:10] [SPEAKER_00]: People, people love it.

[00:24:12] [SPEAKER_00]: Yeah.

[00:24:12] [SPEAKER_00]: And there's this, I think in the call center in Singapore, we've got 500

[00:24:14] [SPEAKER_00]: people on it, which is great.

[00:24:16] [SPEAKER_00]: And so that's great.

[00:24:17] [SPEAKER_00]: So that's one, one side of the core.

[00:24:18] [SPEAKER_00]: The other side of the core worker is if it takes away significant amount of

[00:24:22] [SPEAKER_00]: work, which was administrative in nature, right?

[00:24:25] [SPEAKER_00]: Then we need to think about upskilling.

[00:24:27] [SPEAKER_00]: Right.

[00:24:27] [SPEAKER_00]: And so if you've got an assistant salesperson, how could we make this

[00:24:32] [SPEAKER_00]: person the salesperson, right?

[00:24:34] [SPEAKER_00]: And do a skills analysis and say, okay, here's a program that

[00:24:37] [SPEAKER_00]: will get you out there.

[00:24:39] [SPEAKER_00]: Right.

[00:24:39] [SPEAKER_00]: And the point of this is if you've got more salespeople out there,

[00:24:42] [SPEAKER_00]: you're going to get more sales.

[00:24:43] [SPEAKER_00]: Right.

[00:24:43] [SPEAKER_00]: And so there's nothing wrong in that.

[00:24:45] [SPEAKER_00]: That's great.

[00:24:46] [SPEAKER_00]: Yeah.

[00:24:46] [SPEAKER_00]: So, so that is, there's a level of change management there.

[00:24:50] [SPEAKER_00]: Right.

[00:24:50] [SPEAKER_00]: But it's more a positive story if you like.

[00:24:53] [SPEAKER_00]: Right.

[00:24:53] [SPEAKER_00]: The third, which is where there's a significant impact on the role.

[00:24:57] [SPEAKER_00]: And if you took, we haven't really done this work here, but as an

[00:25:01] [SPEAKER_00]: example, if you took an analyst, somebody who was, you know, whose job

[00:25:04] [SPEAKER_00]: was to pick four, five, six documents every day, understand what they say

[00:25:08] [SPEAKER_00]: to then synthesize it into one, one pager or two pager and all that.

[00:25:12] [SPEAKER_00]: Generally, AI will take these tasks and then can do these tasks in

[00:25:16] [SPEAKER_00]: minutes and all that kind of stuff.

[00:25:17] [SPEAKER_00]: So here we really do think about risk killing.

[00:25:20] [SPEAKER_00]: Now, this is not the first time we're thinking about this.

[00:25:23] [SPEAKER_00]: We've been, we embarked on the digital age in 2014.

[00:25:27] [SPEAKER_00]: Right.

[00:25:27] [SPEAKER_00]: And at that time, nobody was talking about digitization

[00:25:30] [SPEAKER_00]: in the financial industry.

[00:25:31] [SPEAKER_00]: Right.

[00:25:31] [SPEAKER_00]: Everybody was happy with what they were doing.

[00:25:33] [SPEAKER_00]: And we were like, we want to go end to end digital front end web,

[00:25:37] [SPEAKER_00]: mobile instant fulfillment, customization, N equals one, all that stuff.

[00:25:41] [SPEAKER_00]: People are like, Oh my God, we're going to lose jobs because

[00:25:43] [SPEAKER_00]: you don't need operations.

[00:25:45] [SPEAKER_00]: You don't need all the servicing is going to be done by, by through

[00:25:49] [SPEAKER_00]: chat and all that.

[00:25:49] [SPEAKER_00]: Then we don't need call centers.

[00:25:51] [SPEAKER_00]: In fact, no jobs went away.

[00:25:54] [SPEAKER_00]: New jobs appeared and people who got impacted right with the more

[00:25:58] [SPEAKER_00]: mundane operational jobs like sorting paper.

[00:26:01] [SPEAKER_00]: Right.

[00:26:01] [SPEAKER_00]: I mean, they got reskilled right to do operational tasks, to become

[00:26:05] [SPEAKER_00]: makers, to become checkers, right?

[00:26:07] [SPEAKER_00]: They became sales assistants.

[00:26:09] [SPEAKER_00]: They became customer service agents, right?

[00:26:11] [SPEAKER_00]: They went and worked in the frontline in the branches.

[00:26:14] [SPEAKER_00]: So we got to approach it the same way.

[00:26:16] [SPEAKER_00]: And what's going to happen is you're going to find that today, everybody

[00:26:19] [SPEAKER_00]: thinks that it's 90%, it's going to take jobs away and 10% good help.

[00:26:24] [SPEAKER_00]: Actually in reality, it's the other way around.

[00:26:26] [SPEAKER_00]: It always has been when, when, when the cars came, this is what

[00:26:29] [SPEAKER_00]: they thought about the horses actually, you know, it's, it's riding

[00:26:31] [SPEAKER_00]: on the digital wave and it's a big step forward, right?

[00:26:35] [SPEAKER_00]: But it's a step forward and not a complete paradigm shift.

[00:26:39] [SPEAKER_01]: 100% with you.

[00:26:40] [SPEAKER_01]: And I think, as you said there, that reskilling and upskilling,

[00:26:43] [SPEAKER_01]: that is so important.

[00:26:45] [SPEAKER_01]: So it's great to hear how you're pushing that forward too.

[00:26:48] [SPEAKER_01]: And if we look ahead into 2025 and beyond, I know it's almost

[00:26:52] [SPEAKER_01]: impossible to predict the future now, but how do you envision the

[00:26:56] [SPEAKER_01]: future of AI and banking, particularly in terms of customer

[00:26:59] [SPEAKER_01]: experience, operational efficiency?

[00:27:02] [SPEAKER_01]: How do you see all this evolving?

[00:27:05] [SPEAKER_00]: So like I say, this is going to be any trip.

[00:27:07] [SPEAKER_00]: So one, I agree that there will be impact and there will be

[00:27:11] [SPEAKER_00]: efficiencies across many areas.

[00:27:14] [SPEAKER_00]: But this is what a great opportunity.

[00:27:16] [SPEAKER_00]: This gives you the ability to invest in the right places in a very

[00:27:20] [SPEAKER_00]: meaningful and thoughtful way, right?

[00:27:22] [SPEAKER_00]: Like every organization, everybody's got pressure, right?

[00:27:25] [SPEAKER_00]: We've got cost pressures, revenue pressures, market pressures,

[00:27:28] [SPEAKER_00]: all that kind of stuff.

[00:27:29] [SPEAKER_00]: Same everywhere.

[00:27:30] [SPEAKER_00]: This gives us the opportunity to be able to create capacity.

[00:27:35] [SPEAKER_00]: Right.

[00:27:35] [SPEAKER_00]: And with this capacity, how do we invest this capacity?

[00:27:38] [SPEAKER_00]: Now, in some cases, we will want to increase customer experience.

[00:27:43] [SPEAKER_00]: We will want to increase revenue.

[00:27:44] [SPEAKER_00]: So we want to put those resources that we've been able to free up there.

[00:27:48] [SPEAKER_00]: In some cases, we may say, Hey, look, this efficiency is so good.

[00:27:52] [SPEAKER_00]: Right.

[00:27:52] [SPEAKER_00]: And there isn't that much work here.

[00:27:54] [SPEAKER_00]: So maybe we'll take a little bit of the productivity gains there and

[00:27:58] [SPEAKER_00]: reduce some costs in some places.

[00:28:00] [SPEAKER_00]: It's not a, hey, here's a strategy that's going to reduce our

[00:28:04] [SPEAKER_00]: cost by this much.

[00:28:05] [SPEAKER_00]: Well, here's a strategy that's going to increase revenue or

[00:28:09] [SPEAKER_00]: customer experience by so much.

[00:28:10] [SPEAKER_00]: It's a combination of things that one organization is not a single

[00:28:14] [SPEAKER_00]: unit, right?

[00:28:14] [SPEAKER_00]: It's not a single cell animal.

[00:28:16] [SPEAKER_00]: It's quite complex.

[00:28:18] [SPEAKER_00]: And so I think it's horses for horses, but for me, and at

[00:28:21] [SPEAKER_00]: least for DBS, what we look at is this as it's an amazing

[00:28:25] [SPEAKER_00]: opportunity to invest in the right place, dial up custom experience.

[00:28:29] [SPEAKER_00]: Cause if we do that, right.

[00:28:30] [SPEAKER_00]: And we increase customer loyalty, our customers will be sicker.

[00:28:34] [SPEAKER_00]: They will be happier.

[00:28:34] [SPEAKER_00]: They will spend more with us and not leave us.

[00:28:38] [SPEAKER_00]: And that is already a big plus.

[00:28:41] [SPEAKER_01]: And I think that is a beautiful moment to end on today, but before

[00:28:45] [SPEAKER_01]: I do let you go, we've talked a lot about the future where we're

[00:28:48] [SPEAKER_01]: heading and the problems we're solving with technology and you've

[00:28:51] [SPEAKER_01]: had a hugely successful career.

[00:28:53] [SPEAKER_01]: And I'd love for you to look back for a moment and have a think

[00:28:56] [SPEAKER_01]: about someone that maybe helped you get you where you are today.

[00:29:00] [SPEAKER_01]: Someone you're grateful towards.

[00:29:01] [SPEAKER_01]: Maybe someone that invested a little time in you and helped you get you

[00:29:05] [SPEAKER_01]: where you are today, who would that person be and why it'd be great

[00:29:08] [SPEAKER_01]: to give them a little shout out to them.

[00:29:10] [SPEAKER_00]: Oh, well look, there've been different people at different

[00:29:13] [SPEAKER_00]: stages of my career, but maybe the first person that really had an

[00:29:19] [SPEAKER_00]: impact on me was a guy called Andrew Ruan when I was working back

[00:29:23] [SPEAKER_00]: in South Africa and I was doing sales engineering at the time and

[00:29:28] [SPEAKER_00]: we walked into quite a large coal plant outside of Johannesburg and he said,

[00:29:34] [SPEAKER_00]: okay, over to you.

[00:29:35] [SPEAKER_00]: It's your big sale.

[00:29:37] [SPEAKER_00]: You do this then your career is taken off and you're going to get

[00:29:40] [SPEAKER_00]: a promotion and all that kind of stuff.

[00:29:42] [SPEAKER_00]: And I'm super excited, of course, and I went in and talked and

[00:29:46] [SPEAKER_00]: talked and as a sales guy would in their early twenties and walked out.

[00:29:52] [SPEAKER_00]: And that's why I said, Hey, I think we got the deal.

[00:29:54] [SPEAKER_00]: And he goes, I'm not quite sure about it.

[00:29:56] [SPEAKER_00]: I said, what do you mean?

[00:29:56] [SPEAKER_00]: I think I covered every angle possible.

[00:29:59] [SPEAKER_00]: And he goes, here's one lesson for you.

[00:30:02] [SPEAKER_00]: Don't ever forget.

[00:30:03] [SPEAKER_00]: Right?

[00:30:04] [SPEAKER_00]: He goes, it's very, it's easy to talk yourself into a deal.

[00:30:09] [SPEAKER_00]: Right?

[00:30:10] [SPEAKER_00]: It's even easier to talk yourself out of the deal.

[00:30:13] [SPEAKER_00]: Remember when to shut up.

[00:30:15] [SPEAKER_00]: So that, I mean, that was the first, but I I've been blessed.

[00:30:18] [SPEAKER_00]: Actually, I've had many, many people that have been extremely

[00:30:22] [SPEAKER_00]: influential in my career over time.

[00:30:24] [SPEAKER_00]: And I've switched careers between logistics and now in banking.

[00:30:29] [SPEAKER_00]: And you can imagine, you know, in mid career to move industries entirely.

[00:30:35] [SPEAKER_00]: It's quite difficult.

[00:30:37] [SPEAKER_00]: And if you don't have the right people around you to help you and

[00:30:40] [SPEAKER_00]: mentor you, it's quite difficult.

[00:30:41] [SPEAKER_00]: So I'm blessed in that way, but there's one, if you like, so it's

[00:30:46] [SPEAKER_00]: a very important door when to keep quiet.

[00:30:48] [SPEAKER_01]: Yeah.

[00:30:49] [SPEAKER_01]: Great story.

[00:30:50] [SPEAKER_01]: Great message behind it as well.

[00:30:52] [SPEAKER_01]: We've covered so much today in our conversation.

[00:30:56] [SPEAKER_01]: I'm sure there's a few light bulb moments going off.

[00:30:58] [SPEAKER_01]: People are going to want to find out more information.

[00:31:00] [SPEAKER_01]: So for those people listening, where's the best place to find you or your

[00:31:03] [SPEAKER_01]: team online and just ultimately find out more about anything we discussed today.

[00:31:08] [SPEAKER_00]: So obviously I go to my LinkedIn page.

[00:31:11] [SPEAKER_00]: I post quite regularly.

[00:31:12] [SPEAKER_00]: So there's a lot of stuff in terms of where we do keep out, keep a look

[00:31:16] [SPEAKER_00]: out for the press releases that we have, there's many more coming

[00:31:19] [SPEAKER_00]: and a lot of exciting stuff.

[00:31:21] [SPEAKER_00]: We have a DBS LinkedIn page, which is also quite active and our website also.

[00:31:26] [SPEAKER_00]: So plenty of places to reach out if anybody wants to reach out to me

[00:31:29] [SPEAKER_00]: personally, obviously LinkedIn's the easiest way to do it, but there's

[00:31:32] [SPEAKER_00]: a lot of, and not just me, by the way, my colleagues, my team members,

[00:31:36] [SPEAKER_00]: they're also quite active.

[00:31:37] [SPEAKER_00]: They post in various areas, be it responsible AI, be it the use of

[00:31:41] [SPEAKER_00]: tooling platforms, all that kind of stuff, governance, and of course

[00:31:45] [SPEAKER_00]: use cases and what we're doing with both generative AI and traditional AI.

[00:31:49] [SPEAKER_01]: Awesome.

[00:31:50] [SPEAKER_01]: Well, I'll get links added to everything you mentioned there so people can find

[00:31:54] [SPEAKER_01]: you nice and easily and just look chatting with you today about the value

[00:31:58] [SPEAKER_01]: creation that GenAI can deliver and the steps needed to unlock the

[00:32:02] [SPEAKER_01]: technology's fullest potential, not to mention exploring the role of

[00:32:06] [SPEAKER_01]: data stewards, how that role is evolving and ensuring accuracy,

[00:32:09] [SPEAKER_01]: quality and security, and also how DBS, how you're leveraging GenAI

[00:32:14] [SPEAKER_01]: to contribute to things like productivity gains throughout the

[00:32:17] [SPEAKER_01]: organization for everything from systems, customers and employees alike.

[00:32:22] [SPEAKER_01]: We hear a lot about AI, but I think it's stories like that and the

[00:32:25] [SPEAKER_01]: messages that you've delivered today, I think that is what brings it

[00:32:28] [SPEAKER_01]: all to life and what we're talking about here.

[00:32:30] [SPEAKER_01]: So thanks for sharing your insights today.

[00:32:33] [SPEAKER_00]: And thank you very much for having me.

[00:32:35] [SPEAKER_00]: It was good.

[00:32:36] [SPEAKER_00]: Got some nice deep probing questions.

[00:32:37] [SPEAKER_00]: So hopefully people will find this valuable.

[00:32:39] [SPEAKER_00]: A pleasure.

[00:32:40] [SPEAKER_01]: As generative AI continues to reshape the financial industry, how can

[00:32:45] [SPEAKER_01]: banks like DBS strike the right balance between innovation and responsibility?

[00:32:52] [SPEAKER_01]: Well, Nimish has shared insights into DBS banks approach to leveraging

[00:32:56] [SPEAKER_01]: AI for transformative gains.

[00:32:58] [SPEAKER_01]: But most importantly, they're doing this while ensuring ethical

[00:33:01] [SPEAKER_01]: oversight and human involvement.

[00:33:05] [SPEAKER_01]: And the big question for you listening is how are you or your

[00:33:08] [SPEAKER_01]: organization going to adapt similarly in these advancements?

[00:33:12] [SPEAKER_01]: And what steps will you take to unlock the full potential of AI in a

[00:33:17] [SPEAKER_01]: responsible and impactful way?

[00:33:21] [SPEAKER_01]: You've heard from me.

[00:33:22] [SPEAKER_01]: You've heard from Nimish.

[00:33:23] [SPEAKER_01]: What about yourself?

[00:33:24] [SPEAKER_01]: Email me techblogwriteratlook.com, Twitter, LinkedIn, Instagram, just

[00:33:29] [SPEAKER_01]: at Neil C Hughes.

[00:33:30] [SPEAKER_01]: Let me know your thoughts and how you're tackling these topics.

[00:33:35] [SPEAKER_01]: But that's it for today.

[00:33:36] [SPEAKER_01]: I've got another guest lined up bright and early for you tomorrow.

[00:33:39] [SPEAKER_01]: We're going to go into a completely different topic, but more than

[00:33:41] [SPEAKER_01]: anything, thank you for listening today and I will speak to you all

[00:33:45] [SPEAKER_01]: bright and early tomorrow morning.

[00:33:47] [SPEAKER_01]: Bye for now.