Data management experts discussed the value of data and how financial organisations are leveraging this key business asset by embedding it in design. Key challenges related to data management and governance such as ensuring data accessibility and overall quality require a collaborative approach within and beyond the organisation to effectively resolve.

Céline Le Cotonnec, Bank of Singapore; Monica Srinivasan, Standard Chartered Bank; Adeline Kim, Visa; Yongnien Tan, Cagamas; Robert Weir, formerly of CIMB and Peter Ku, Informatica shared practical insights on creating a data culture and leveraging fit for business data in financial organisations to drive key business imperatives.

Cotonnec highlighted the need to focus on data literacy and culture in the entire organisation and not only in specific departments related to data management. She urged organisations to rethink the role of data and automation in their internal culture. “We need to think of how we democratise data within the organisation—by leveraging more advanced analytics, such as AI techniques, and the overall data engineering framework and building data models,” she said.

Srinivasan commented on the risks pertinent to data management and the implementation of holistic means to manage valuable data. She also emphasised the importance of practising data embedded design by financial institutions. “We must operationalise and practise data by design. Thereafter, for every data-heavy project, new product, business or initiative, it is imperative to assess if data is fit for purpose and having those ongoing controls.”

Kim discussed the customer experience as a centre of data management, insisting on driving data to offer a better customer experience. Meanwhile, Tan shared his observation on the lack of data standardisation in the mortgage industry.

Weir focused on how organisations need to bridge the gap between the operational capabilities to leverage data in their business models. “There is a gap between access to advanced technologies and the ability to exploit those technologies because even the technologists have a little learning to do to make it right.”

Finally, Ku highlighted the edge that the latest AI, ML and cloud-enabled technologies have given financial organisations to automate and transform their data management processes. He emphasised the importance of considering data within organisation as a business asset to be protected and managed properly.

The following key points were discussed:

• Running financial services organisations as data businesses can be achieved by considering data as a business asset.
• Understanding data and processes is crucial, but also creating the structure, talent and skill set to manage data is an ongoing challenge.
• Advanced AI, ML and cloud-based technologies vastly automate and transform the process of ensuring data quality.
• Managing data quality is one of the key challenges faced by institutions that requires a collaborative approach to eliminate unfit data in the business.
• Leveraging the cloud to access and manage data helps businesses reduce time-to-market and address data-related challenges.
• Identifying clear objectives to utilise and classify data to align with the organisation’s strategies and ensure a customer-centric approach.

The edited transcript of the session:

Foo Boon Ping (FBP): Good morning and welcome to The Asian Banker Radio Finance, I hope everyone is having a great start today. We'll be discussing how financial services organisations unleashing the power of data to turn themselves into data businesses by fully leveraging the most important assets that they have, data. With us, we have a group of experts whose business is managing data for some of the leading financial institutions in Asia and the world. We would like our audience who have joined us via Zoom, Facebook or LinkedIn to be part of this conversation. As you share with us your questions and comments, simply type them into the comment section of the respective platforms. They'll be relayed to me and I will in turn post them to our guests.

Let me introduce them now. I'm very excited and happy to have with us this morning experts in the data management field to share their experiences and insights starting with Céline Le Cotonnec,  chief data innovation officer at Bank of Singapore, the private banking arm of Singapore-based OCBC. She oversees the data strategy of the bank and the implementation of data science and analytics use cases in cooperation with OCBC’s group customer analytics and decisioning, as well as its AI Lab.

We are very happy to have also Monica Srinivasan. She's the executive director, responsible for data management of global bank, Standard Chartered. She provides ongoing assurance and compliance to data quality management framework for corporate and institutional banking clients, as well as financial client crime, and compliance processes.

Next, we have Tan Yong Nien, senior vice president of technology and operations at Cagamas, Malaysia's mortgage corporation, the equivalent of Fanny  Mae in the US. He provides operational and strategic leadership in technology and operations as a mortgage corporation.

Next, we have Adeline Kim, head of data solutions for Asia Pacific at global payment company Visa, where she leads the open banking risks and other data-related solutions and functions.

We also have Robert Weir, former programme director of data at Malaysia-based CIMB Bank. He has led the implementation of multiple master data management workstreams in Malaysia, as well as in Singapore.

Finally, we're very pleased to have Peter Ku. He is the vice president and chief industry strategist for banking, financial services and insurance for Informatica. He is responsible for defining Informatica’s Go-to-Market Sales, Marketing and Partnership strategy to help the industry manage, govern and protect data as a business asset, leveraging Informatica's Intelligent Data Management Cloud platform.

Welcome, everyone and thank you for being part of our panel of experts. I look forward to a very delightful and insightful discussion with you on a topic that I'm sure is close to many of our hearts, at the top of mind.

Data management and governance, essentially what you do everyday to get your data, and organisations ready to support all the business imperative to grow and to sustain your operations across all aspects of your organisations from how you communicate to and win a potential customer, to safely onboarding that customer and delivering him or her a highly personalised experience in order to increase sales, share of wallet to manage risk, ensure regulatory compliance, and to prevent financial crime and fraud. Your roles are all encompassing. Data as a subject and data management as a concept is familiar to all of you, especially now, in what we hope is a post-COVID work that has already witnessed accelerated digital transformation, and what we believe is an irreversible shift to virtual operations and transactions.

Data is at the core of this wave of digital transformation and innovation. However, transforming incumbent financial organisations into data-centric businesses is my small complex and is as diverse and varied as the many financial institutions that exist and some of whom are represented by you today. Each data challenge indicated, mainly by organisational and system in new silos that hinder data availability, accessibility and overall data quality is unique, long-standing and often ongoing. There are no simple solutions and quick fixes to them. They require holistic and continuous configuration and reconfiguration of foundational elements to ensure success. It starts at the very calm, the highest leadership level and pervades throughout the organisation, across structure, culture, people, talent, process, system and technology. Nevertheless, the financial industry is making progress in overcoming these challenges.

Some of you have taken a list of internet and technology companies that are data businesses, and they are called to become data businesses yourself. You are realising the real value of data by adopting data culture supporting data literacy, and enabling data responsibility and accountability. You are embracing internal as well as external collaboration  to overcome your data-related challenges. Technologies, as we said, are also fundamentally changing how you manage and leverage data. For example, the proliferation of application programming interfaces (APIs) and cloud banking are helping overcome the limitation of legacy silos system by enabling greater connectivity, agility and scale and the use of artificial intelligence (AI) and machine learning (ML) is also helping to improve the very quality of data. We will discuss the significance of how you, as financial organisations are transforming into data businesses and the benefits of utilising fit for business use data in your business and business models. We also delve into the role that non-industry players can play to help financial organisation deal with your data-related challenges.

I would like to start by asking each of you to briefly describe what you do in relation to data management and to share just one observation of what you see as the current state of data in the industry today from the example of perspective of your own organisation.

Let us begin with Celine from Bank of Singapore. You see it in Singapore where the regulator Monetary Authority of Singapore (MAS) is driving a lot of data-related initiatives, from regulatory reporting, having common data, digital data reporting, to open banking SGFinDex, for example. Tell us from your perspective, what you do as well as how you see data being leveraged in the industry today.

Céline Le Cotonnec (CLC): Thank you very much for the opportunity to talk today.  My name is Céline Le Cotonnec . I'm the chief data and innovation officer of Bank of Singapore, which is the private banking arm of OCBC Bank here in Singapore. My scope covers data governance. How do we manage data? How do we protect data? How do we ensure its quality? How do we democratise data within the organisation, by leveraging on analytics, more advanced analytics, such as artificial intelligence (AI) techniques, and the overall data engineering framework and building data models?  So, it's a kind of  360 view of the data going from, the second line of defense type of activity, to a more first line of defense to what are the insights that we're generating.

In Bank of Singapore, we have taken the stand to democratise data within the organisation. We want to make data available for business decision, swift business decision, as soon as possible. Whatever  your question is, that is related to the business in this world that is currently volatile market, being a global organisation, we need to make insight available as quickly as possible to our business user, risk function, compliance function, so as to be aligned into one source of truth.

Financial organisations have always been working with data. The difference now is the velocity, the amount of data, as well as the technology that is available, also to be able to do data wrangling, insights and generation on a quicker basis. The most important thing we often focus on when we talk about data is the platform,  We need to have the right tech stack and the capability is tremendously important. But I would say this is only 20% of the problem. The main thing that I would say that organisation usually don't focus enough is on the question of data literacy, data culture, and the people.

Here, I would encompass the trainings that are needed, but also, how will data automation and AI, change the role, and the job that your people have today in your organisation,  because it will. so it needs to be as well worked out. In terms of future workforce, what are the required skill sets, but then what will be the role of some of your people today in operation or  even in product development or risk management?

FBP: Okay, so getting that greater data literacy across our organisation, but it’s starting, and a lot of organisations are doing a lot of work there.  Let's hear from Peter. Peter works with many organisations on their data management challenges across the industry, as well as across your geographical location.

Peter Ku (PK): Well, thank you. Thank you everyone for spending time with us today out of your busy schedules. So I'm dialing in from Northern California here in the US, it's a little past 7pm. The day is just getting started for me here, but just briefly, Informatica. We provide solutions to help organisations to leverage and achieve data that's fit for business use. What we mean by that is helping organisations access the data that they need, govern it, protect it, manage the quality requirements that they need to address, as well as making data transparent, so that they can answer questions such as, where does all this data come from?, where does it exist? How was it used? We have a solution platform that allows organisations to leverage these capabilities individually, but also be able to leverage it as a comprehensive solution. We've been in business for over 27 years. I've been with the company for almost 15 years.

Running financial services organisations as data businesses can be achieved by considering data as a business asset

I  came out of the card and mortgage technology space. I spent time spent some time at Visa and other organisations in the issuing side. I tell you, my observation in the industry is this. There's a lot of changes that's been happening in financial services. The adoption of cloud computing has accelerated over the last 10 years. AI and machine learning (ML) had become household names. Data governance is no longer a nice to have, and data is no longer a byproduct of a transaction or interaction. It's a business asset that has to be protected, managed and leverage. And the custodians of data are not just in back office, IT but it extends all the way into the front office.  So the ownership of data is shared by everyone. One thing that I am seeing, however, in light of all the investments organisations are making to modernise and digitise their business capabilities. We also see a significant transformation and modernisation effort in the areas of data management, data governance, and data privacy protection. These are areas where Informatica has been providing solutions for the time that we've been in business.

FBP: Okay, great. Thank you, Peter, for that insight, and next, we will hear from Monica, who's running data management for Standard Chartered across various institutions and corporate business clients, as well as financial crime.

Monica Srinivasan  (MS):  Thank you, Boon Ping. Thank you for having me here, looking forward to contributing and also learning from each other. I am Monica Srinivasan.  I'm the executive director of SME for data quality at Standard Chartered. I'm sure you all are data veterans out here, but I always start with the basics. I would like to give you all a very quick refresher of how, when and why the very basic concept of data quality got its place today.

The US subprime mortgage crisis almost a decade ago that contributed to global financial crisis eventually resulted in Basel Committee on Banking Supervision (BCBS) regulations back in 2013, issued by the Basel Committee, and it completely changed the way that banks looked at data quality. Hence, data quality as a concept is probably the oldest in the family of the data at risk. Despite that, there is a need for continued focus and investment on data to ensure that the data is fit for purpose because it's definitely not a one-time activity.

Where are we today, and what is the one thing that I'm focused on? With all digitisation journey heading towards the peak, there is more data giving us great power. With great power comes great responsibility, isn't it? We must make sure the risks are mitigated and managed holistically. The data plays an integral role in AI, privacy, data sovereignty, and not underestimate records management is also heavy on data and has direct relevance to anti-money laundering (AML) regulations. Our current focus is not just to bring together all the data-related risk for effective risk management and compliance with regulations, but also to use data in the safest manner to drive better value and outcomes for both clients and business. We shall discuss more details gradually. Thank you.

FBP: We'll go deeper into that,   the use of data for risk. Next, we have Adeline from Visa. Visa is working on data solutions for its member financial organisations. Let's hear from Adeline.

Adeline Kim (AK): Thanks Boon Ping. Hi, everyone. Good morning. My name is Adeline and I lead data solutions in the region. In the preparation leading up to this session, I was kind of half-joking with the other panellists, and I'm probably maybe young and one of the ones with a bit of a business lens. So pardon me, especially Monica, as I look into the future,  and remain very optimistic about what data could bring to us. We need to manage everything that you mentioned. In my role, I manage a suite of data-driven products that help our clients drive security, which is really important in the payment network, authentication high approval rates and just broadly, decisioning, so that at the end of the day, what we want to achieve is a better consumer experience like seamless and frictionless. A thing we can't forget,  consumers as a centre for everything that we do around data.

Boon Ping mentioned, the pandemic has definitely accelerated digital migration for businesses, consumers, and in Visa, we believe that this trend will only continue to grow. This explosion of digital payments, we believe offers vast opportunities for business growth. Many of our banking clients are really contemplating how to evolve their business, including our partners and some of the partner banks on this call. In particular, open banking or open data is of interest to us. I use it interchangeably, but I would like to highlight that what we see beyond the regulatory compliance requirements is  data flows. That is what open banking will eventually drive, the whole concept of open finance.

We think that although application programming interface (API) is still evolving and in Europe its really established. In the coming years, there will be immense growth opportunities. Maybe I'll just zoom into,  two or three benefits that we believe will come out of open data. There will be more data sharing and data collaboration within banks and other players in open banking ecosystem. This will mean that financial institutions, and also these players can think about how to …I'm trying to avoid using the word monetise, but I guess it's easily understood.  I'm trying to avoid it only because sometimes it's perceived as a dirty word. How to optimise data access and offer and look at how complementary datasets can be combined to develop better products for the end consumer. Bank lending is one good example, credit lending and credit access is something that every government wants to solve. This will be a good use case.

FBP: We'll go deeper into this as we go open data, the increasing collaboration, the building of ecosystem, which is a point that Tan Yong Nien can talk about because he works with partner financial organisations that provide mortgage services, and from his perspective, running a mortgage corporation, Yong Nien.

Tan Yong Nien (TYN): Good morning, and thank you Boon Ping for the introduction and for having me in today's programme. I’m Tan Yong Nien from Cagamas, the Malaysian National Mortgage Corporation in the secondary mortgage industry. What I do is slightly different from the rest of the panel experts on data today. In Cagamas, I'm both from the technology, as well as the business operation side of the business. In technology,  my teams manage the data platform, as well as data management and governance. On the operation side, and this is where it gets more interesting for me, my team is both the custodians of Cagamas data, and also the user of the data to perform their day to day job in administering and servicing the secondary mortgages. My one observation, Boon Ping, if I'm allowed just to make one would be, if I take the mortgage industry lens like what Adeline mentioned earlier. It will be the data standards or the lack of uniform data standards in the Malaysian mortgage industry for exchanging data between organisations. This would then provide a common vocabulary and taxonomy for the mortgages industry and facilitates the exchange of information across multiple stakeholders, including the lenders, borrowers, services, insurance, as well as regulators. The challenges of data management within organisations, while difficult, are better understood compared to the challenges of data standardisation and exchanges between organisations because it involves, multiple parties and agreements between those parties. That will be one of the key ingredients, for the mortgage industry to be digital, Boon Ping.

FBP: To become digital, to come out with a common set of standards that are most often more usually easier handled by a regulator or an industry association like what's happening in some jurisdictions, like Singapore, for example. Finally, let's also hear from Robert Weir from CIMB in Malaysia, who runs data, master data workstream, and to give us his take on his observation of data in the industry today.

Robert Weir (RW): Thank you. Good morning. Yes, I'm Robert Weir. Until very recently, I was the programme director for data at CIMB Bank in Malaysia. Over the years, I have worked for a number of banks in the UK, Malaysia and Singapore. I've been in data management space for quite a long time, starting when data warehousing just jumped onto the scene. What I've observed over that period of time is the businesses want from data hasn't changed a great deal between in that period of time, which is around  maybe 2000. The banks' businesses want better relationships with their customers. They want to sell, be better at exploiting data as a commercial asset. In that intervening period, we're still dealing with a lot of the issues that haven't been resolved. The technologies as I've experienced with master data management - those technologies have become very sophisticated and capable, and very interesting  to deal with and they can solve problems, but we're still not in my view, anyway and it does depends on the maturity level of an organisation, but data maturity level. The governance aspect of it still needs work, we still need to embed processes to essentially industrialise what we need to do with data. We do some things well, we don't do other things so well. It still takes a lot of effort,  especially in the data quality space. We could build the tools and the interest capability out there that could make the data management space much more closer to realising thebenefits that the business perceives of data because there is an appetite in the business to realise data as a business asset. But we need to bridge that gap between operational capability to realise the capacity and the skill that your tools can provide.

FBP:  Okay, great. Thank you, Rob. That is a very good segue into our next topic for discussion. So the aspiration to be data businesses is to leverage data as a business asset. All our panellists agreed on that point.  What point and what are the gaps in realising such  aspiration? Why do financial services organisation need to be run as data businesses? Our main theme for the discussion today. I'd like to get a perspective from the rest of you, just as Rob has given, is that proposition, all financial organisations run as data businesses. Where do we think we are today? Give us an insight as to your answer.  I'm sure the answer is yes. Every organisation wants to be run as a data business. Where are we now? Where are the gaps? Where are the challenges? How have you defined your own priorities agenda around that? We're going to go through the same lineup, all the panellists in your response, starting with Céline.

CLC: When it comes to our data strategy,  involving the rest of the organisation and trying to have,  an enterprise architecture that is  feeding  our business needs, I would say that  data is used differently, according to what is your business function. What are your role and responsibility. But one thing that is common is that we need to have a common definition of our data fields and then look at  how do we have common data model. Data is also a language, we need to speak all the same language. So starting by defining clearly what are the data, the metrics that the organisation wants to rely on in order to move the business, the way the strategy of the management committee, then,  it’s the first step into creating an informational data-driven organisation.

FBP: Earlier, you mentioned, you're working a lot on data literacy and democratising data so that the use data goes to the end-user for provisioning for all aspects of their work. How has that journey been in developing data literacy throughout the organisation? And making sure the skill sets are there? There is a common understanding of the importance of data across the organisation.

CLC: It's a question of target operating model, right? Like you, you tend to have a data team. I ran the central data office for Bank of Singapore, which we call the data hub. We are the centre of expertise.

Understanding data and processes is crucial, but also creating the structure, talent and skill set to manage data is an ongoing challenge

We are also implementing a small analytical team in each department, so we have one in risk, one in compliance, one in product, one in finance, one in human resource (HR), one in operation, as well, who has a lot of need of data for efficiency, purpose, marketing and digital channel.

We have those analysts that are  spread over the business. They all have Python, structured query language (SQL) and visualisation skill sets, so they themselves will support the department. And then we have a massive upscaling programme that had been planned with HR, so we are teaching people how to use SQL. We’re trying to, I wouldn't say get rid of Excel. It  will never be possible, but at least everything that is today sent through Excel by people in the organisation using email, for example, we're trying to put that into a common database, and having everything shared and made accessible, the Coda  share.

We often say that IT needs to understand business. But the other way around is also right. Business needs to understand how IT processes work. What is the deployment process, how do you document a code,  and they need to go from expressing requirements using Word document or whatever the support is, to  being able to code whatever they want. Because they are the ones that have the business knowledge. If they're able to translate that into a code, then  the overall organisation would benefit. So we have a massive upscaling programme. We have a data certification pathway. There's a local university that lasts for nine months, three days a month, and we have already 200 people that have gone through that certification. We have an e-learning partner. It's a constant effort and it's also in the key performance indicator (KPI) of the people to upskill themselves.

FBP: How long has this started and give us just a glimpse into how many of the department within the bank itself has got this data skilled analyst and beta from them? Is it 30%, 40% 100%?

CLC: No, I would have  2,000 people in the bank. We have 215 people that are today available to  add the ability to create a dashboard and then do SQL queries. It's about 10% of our population. We are extending to the front office now, it was mainly for the support function, originally. We have about 60 people in the organisation that are pro-efficient in Python to automate their own tasks. The goal is that we are creating monitoring between the more advanced Python developer and then the rest of the organisation, so everyone is able to  automate their own task.

FBP: Okay, great. Thanks. Next, Adeline, give us a perspective of how grady is the member or partner institutions that you work with to deliver solutions that you hope to work with them on.

AK: Thank you Boon Ping.  So my answer  to your question is, yes. Why?  We all know that with digital commerce and the explosion, we're all dealing with a different set of competitors. The traditional competition set that we used to deal with is no longer enough  for us to watch out for. We see this with our FY partners, and even  within these, that's what we're seeing as well. So for example, when banks are no longer dealing with other banks, some of the competitions that are dealing with are pure data businesses.  That's why you tee up this question.

The partner banks that we work with we’re thinking how they should be structuring themselves. I’d like to maybe offer a thought,  that maybe it's not a binary thing. It's not like,  you need to be either non-data business  or data business is probably an evolution. Within Visa, that's kind of what we went through. So we didn't  have to pivot in a way. What we did was,   if you think about the origin of Visa, we changed, we help digitise money. So, physical money became digital money. And we've evolved that thought to think about  what sits behind digital money is really tons of data. An online transaction that you make today derives a lot of data around where you are,  what device you're trying to transact on -- is it a mobile device, is it on your laptop, where you transact, what currency, and the data that we collect then allows us to do a lot more, either by ourselves or with a partner banks. Just sharing the Visa journey, we're now moving towards or evolving our thinking towards… we're now moving  our thinking from money movement, which is what we do very well into data movement. In particular, within the data movement space, what can we do to develop more value-added services that can help our partner clients, and optimise the massive amount of data that they collect?

Consumers are becoming very astute. What we want as consumers? We want our consumer commerce experience to be secure, reliable, protected,  but also personalised. This is where data becomes important,and when used correctly, the data can power not just a secure commerce experience, but also drive greater customer engagement. In order to achieve this, we know in Visa, we can't do this alone, and  our partner banks acknowledge the same thing as well. We have this term called ‘Network Network’, which is really- we are out in ourselves, but we are partnering with other networks that power data  that helps move money and move data. This is something that we see our bank partners, and not just bank partners but fintech partners that we're working with, do as well.

FBP: Okay, so you're working with partners across that solution and also enabling them with the solutions. Thank you, Adeline. Next,  from Yong Nien your perspective on financial services as data business.

TYN: It’s a definite yes for me, Boon Ping. In the mortgage industry, if I put on my operations lens are just bits of data moving through a model or workflow. Starting from the lenders’ applications to approval, and that same mortgage has been then packaged with others like it and sold to  secondary mortgages organisations like Cagamas. These are just data moving between organisations. I can give you a multitude of reasons Boon Ping, why I said yes, but the first reason would be, better data always results in better decisions, especially in business operations. The workflow that I mentioned just now, at every decision point of that workflow, the better and the more data that we have, and the better quality of that data, it’s almost always results in better clarity, and eventually better decisions.

Boon Ping, if I can just cite one example in the mortgage industry, in the mortgage loan approval process. If I look at  two aspects of that process, the decision points and the data that support how these decisions are made at those points. First,  on the decision points in a typical mortgage application process we have -  whether a person qualifies for a loan, how much a person qualifies for, and what their interest rate will be? Let's look at the common data sources that support the three decision points that I just mentioned. So these are commonly - a person's credit score, the person's job and income, and how much debt the person has. Finally, probably,  a valuer’s appraisals of the property. What if, we can enrich this data further and gives it more dimensions,  For example, the industry that a person is working at, is it poised for growth? Potentially giving this person a better job, prospect and income in the future?  Like our fellow data scientists, in this panel. More on the property, is it  up-and-coming township based on recent transactional trends? With this additional data, how would then  that those decisions, be outed. Will the bank be willing to take on more risks of lower interest rate, and a higher loan amount, or potentially even contribute to more creative product innovations. To my fellow panellists,  I'm not grossly simplifying the entire mortgage process, but I guess what I'm trying to illustrate here is that, with more data,  that always comes  for a better decision, and ultimately, I believe that, in the mortgage industry, that would then result in better pricing,  for our consumers. Improve efficiency and cost management for industry players, better understanding of risks for insurers, and very importantly,Boon Ping  deeper insights  for our regulators.

FBP: Yes, greater decision for better product, better customer experience and greater regulatory compliance. The challenge is that each institution is at a different stage of being data ready, and how to get everyone onboard. As you mentioned, getting some common standards, right across the industry, especially if the regulators or association, are driving some of those initiatives.

Finally, back to Robert, before we move on to the next topic. We understand that the challenges in operational capability versus the technology potential to use data -- those challenges are there. But there are also challenges with the management approach to using data to look at returns on investment (ROI) on data, for example. Could you share your thoughts with us, Robert, on that?

RW:  Yes, Return or payback for investment in data is a big topic in an organisation or a bank. It's something that I had to deal with quite a lot in a bank. What strikes me about the question is about being data-focused. Banks are data-focused, they always have been  for quite a long time. They've been building risk models for a long time. We've been in this space,  even the  issuing a statement is full of data,  it's there. What is less clear, what I've seen  is that the banks are operating in a very competitive, very tight markets,  so strategies have to change. The link is an area where I'm really quite interested. I would like to see more on the  link between the organisation's strategy. Where does it want to go? Through the development of use cases Would be a very good, strong vehicle for shaping how data is meant to perform. How it's meant to be stored, how it's meant to be collated?

Over a period of time,  when the banking crisis happened, the accusation of banks was that  they had been overly product-focused to sell products. The more products you sell, without the technology, a risk, will have consequences and it did. And then shifts you towards the focus on customer. But the customer itself is a strategy. Dealing with a customer is a strategy and what you want to do with that is know the shift in focus of how do you make data, continually relevant to the business objectives and the strategies of the organisation. Will in a lot of ways help to deal with the question of return?  What am I getting from my money? Because you can relate it more directly to what is happening at an operational level within the business through  very aligned use cases.

FBP: The use of data and how is that aligned to the overall business strategy and objective and that significance, which kind of relate to our next topic. What is the significance of having data that is fit for business use? What are the challenges and the root causes of so-called unfit data? I hope this is quite a straightforward topic, non-contentious questions, but I'll get Peter first to define fit for business use data for us, and the challenges of deriving fit for business use data and root causes of unfit data.

PK:  Sure. So let's take a step back. We hear across the financial services industry, the top business imperatives from improving customer experience to growing the business across cross-sell, upsell, expanding wallet share. We talked about regulatory compliance, that's obviously a fact of life in the financial services industry. Regulations don't go away, just new ones get added on top of the existing ones. You also have the need to strengthen risk management, especially in today's world where we see the effects of climate change, pandemics, geopolitical concerns that are happening in certain regions of the world, and what the financial implications or the implications are to the financial markets. Adeline, you mentioned this,  our competitors are the competitors in the financial services space. The traditional brick and mortars aren't other traditional brick and mortars, you have the fintechs, that are eating the lunch of many of these larger established companies across all sectors of financial services.

Advanced AI, ML and cloud-based technologies vastly automate and transform the process of ensuring data quality

As organisations invest in new digital capabilities, analytics, AI, all the things that we've been talking about earlier, data has to be fit for business use for the people that are being asked to run the businesses today, from the sales folks on the wealth management side of the business to the compliance officers to the people that are combating fraud.

Number one, data has to be accessible. It has to be accessible in the systems and applications that those people use each and everyday. Data comes from multiple systems across your data centres in the cloud. If it's not in those systems that they use everyday, it's useless, those applications are useless. It has to be cleaned. We take this for granted. Data quality problems have existed since computers were created. When those zeros and the ones were turned into round the clock information. Quality sometimes is also in the eye of the beholder. But people need to understand the data that they're dealing with. What they're using to do their jobs is clean and trustworthy.

Number two data has to be valid. You need to have valid merchant IDs and merchant programme codes and you have to be able to say, the data that we use in order to quantify our positions around credit risks, is again valid.The reference data that we're using --valid identifiers, valid addresses and email information. Next, in the world of transparency, data has to be transparent. We need the ability to answer the question, where did they come from? How was it created? Where is it used, and this is in the areas that we're talking about. Here is data lineage, the ability to  trace where data gets created all the way through its lifecycle to when it's consumed.

Next, business, people need to understand what their data means. This is why data governance was so important. There was a lot of data in the organisations that we call financial services today, that they had 20 years ago. What was missing was business context. How is that data used? When are we using it? When should you not use it? What are those critical data elements to your point? Earlier, we're speaking Robert about being able to tie your critical data elements to the activities that generate revenue for the business, as well as the represent cost to the business. Because if you can do that, you can  calculate a metric that most companies can understand, net income per employee. People need to understand what data they have and how they should use it.

Finally, data has to be protected. We live in a world of cyber thieves and  criminals out there trying to steal our data assets in order to take advantage of our customers and our employees. But you can't begin protecting that data unless you know where it's unprotected. What data is unprotected and should it be protected, whether it's data at rest or data in motion. I mentioned these aspects of data being fit for business use. This is summarised -- it has to be accessible, clean, valid, transparent, understood, and protected because  data is not fit for use. I don't care if you’re a mortgage servicer or credit card issuer or life insurance company, or a software company trying to sell technology such as ours, it is not going to mean anything to you. That data becomes useless for the business if they  don't have their data that's ready and available to them. So we talked about all these things, because organisations struggle in addressing their data requirements for two reasons. One, is they've thrown bodies at the problem for many years. There's a lot of smart individuals out there in the world who believe that they can.. It's their job as a data scientist to fix the data problems before they can  run their models and come up with some recommendations. It's  not their job to fix these data issues.

The technologies and the tools that organisations have adopted over the years have  become outdated. In today's world of massive amounts of data, different types of data, the velocity and the exchange of data across on premise into the cloud require modern capabilities. And this is where we see the companies across the globe, from small banks to large insurance companies, now looking at how do we address  data governance, data management capabilities? And the good thing is many organisations are.

FBP: Great. Thank you, Peter, for bringing that perspective on what fit for business use data is and so the challenges of that organisation in achieving that. Monica, could we get your perspective on that, and from what you do in data management, at the same time?

MS: Sure. Thank you, Boon Ping, and thanks to Peter. You talked about the principles behind data quality very well. I'm going to talk about it slightly differently. Data quality is meaningless as a standalone concept. It only makes sense in the context in which data is used. That's what Peter was also insisting. For example, simply saying name is incorrect does not mean anything really. However, if you also say that an incorrect name shall impact our ability to screen names on a watch list for financial crime purposes, it makes absolute sense because you all know, there is a sanctions risk impact. So, likewise, a name being incorrect is less relevant for pure accounting. What I mentioned, here is a very simple use case. But these continue to grow with changing market conditions and growing business needs so we must operationalise and practice data by design. Thereafter, for every data, every project, every change, every new product or business or initiative, it is imperative to assess if data is fit for purpose and have those ongoing controls. So assessment should also include various dimensions like whether the data that is required for that particular change or project is complete, and what is the level of accuracy? All of this should be part of the project plans.

Managing data quality requires a collaborative approach to eliminate unfit data in the business

Now, in challenges and root causes of unfit data. Boon Ping, you said it is straightforward and non-contentious, but I still think data quality as a concept is widely misunderstood, because I'm a believer that data quality issues do not always originate at the source. So there are definitely issues mostly sitting elsewhere, either because we have moved forward, or for whatever reasons.

Let me try and explain it with an example. Let's simply talk about food analogy. So let's say a customer walks into a restaurant and order, ‘Sir, vegetarian soup’, but does not specify spice levels. The server carries the order back into the kitchen and the chef delivers a spicy vegetarian soup. Now, you might have had an expired chicken breast in the kitchen. But that does not really matter because the soup is a vegetarian soup. Or the customer might complain of spice levels, but that was never specified in the first instance. So in both these scenarios, the underlying data quality of the ingredients used in the soup is good. However, it is about handoff between the chef and the ultimate consumer.

This is what happens in the data quality, where often the issue is primarily because of misinterpretation or lack of data handoff between the actual data provider and the end consumer. You can imagine if such issues can happen in a simple setup, with just three parties, in a large organisational setup with numerous systems and innumerable data interfaces, we must remember that there is still a lot of misinterpretation that can creep in. So data in any organisation is rarely created with full awareness of every possible downstream usage of that data. So my theory is where there is a collaborative approach to data handoffs, it can definitely help bring down the perceived data quality issues. That is what I wanted to table it here. Thank you.

FBP: Okay. Thanks, Monica. That’s a very interesting insight. It’s not in the data itself, it’s how the assumption and the context under which that data is being communicated and used. I believe, Celine would like to also add a comment to this.

CLC: I really like the image that Monica was using because it's a simple one, but it definitely describes,  the issue that we can get when we talk about data quality. And just to add on the people and cultural aspects. So there is a collaborative approach that needs to be here. There is also  an accountability. For everyone in the organisation, data is an asset. It's not only the data management office, or it's not only IT or it's not only  the data custodians that are responsible, or the data owner for that matter. It needs to be a cordon, even like a democracy, if you wish. Like everyone is responsible to make that right.

This is where definition, as always, we're coming back to what are the good practice in  data management. When it comes to lineage, being able to know what are the impacts of this specific data field on  the downstream systems or report. And then being able to talk the same language and this is where we go back to data culture and data literacy. We need to call,  a carrot,  a carrot and nothing that a carrot is an apple, for example, right? Sometimes people use the same word, but have a different understanding behind that's also one of the person of data quality, I believe.

FBP: Okay, great. Thank you. We've got a question that has come in from members of the audience. This relates to security, data security, and data quality. Regarding data security, how do you see data security and its relevance to data quality? Do you see any overlap between quality and security? And if yes, can you please explain with example. Céline, maybe would like to give a stand at that? And then maybe we'll ask Monica for her perspective as well, on data security and relevance to quality? I'm sure it is, like in terms of data being needed to be protected.

CLC: So it does. Definitely, it will start with a clear classification of your data. That's usually a tremendous exercise that needs to be done. What do we classify? We have usually some different levels of classification for information security. How does that translate into your structured data within your system in your organisation? What is the level of protection that needs to be put in place for your different types of data sensitiveness? We have personally identifiable information (PII) data  and personal information that are now heavily regulated by Personal Data Protection Act (PDPA) in Singapore, General Data Protection Regulation (GDPR) in Europe, and more countries are  drafting their own data protection law. It makes it quite complex for the bank to be able to  comply with all of the various requirements from the various data privacy law. But if you want to democratise data, you need to be able to understand which one are sensitive and those that you need to make available to the rest of the organisation. The challenge is always to find what is the right balance between, protecting the data while democratising access to the data. At the end of the day, this is where you need to be very clear in which data do you classify as highly sensitive. Usually, when you look at good data management practice, and I only took less than 2% of the data of an organisation and I'm talking about structured data and unstructured data are highly sensitive data. That would be merger and acquisition. That would be  like very highly strategic information and usually, most of the employees don't have access to those data. They are usually unstructured format. Then after you got the question of personal data of your client, that should be tokenised, anonymise, in order to be better protected.

FBP: Okay, great. Thank you Celine. Peter would like to also respond to that question.

Peter Ku (PK): Yeah, to someone's point there, I think data quality, versus data security. What we've seen organisations deal with is number one, we found situations where the data fields were  mislabeled. When we ran through the analysis, we discovered that it wasn't  a phone number of a foreign country. But it was a citizen identifier, which was deemed personally identifiable. Therefore, that was sensitive data that needed to be protected. Securing data depends on what type of data you're securing. And obviously, PII data is going to be treated at a higher level than non-PII information. This is where you have to ensure that you have the right definitions. You have the right naming conventions, that people know what they're protecting. Because you don't necessarily treat and secure all types of data at the same level. So quality does matter at the end of the day. And because there are costs associated with it, this cost is secure and protects sensitive data. If you don't have good quality definitions; defined naming conventions, then you end up wasting precious resources, protecting something that ultimately doesn't have to meet that level of security. That makes sense.

FBP: Great. Thank you, Peter. Yong Nien, would like to respond as well.

TYN: In terms of security, that's given  in our industry,  especially,  financial industry, mortgage industry. We are very well governed, by our regulators. So I see that it goes hand in hand, between quality and security. It's not so much overlap. But, it goes hand in hand with making sure that the data fully reflects what we have. At the end of the day, data is a representation of the actual physical world. The regulators hold us accountable to making sure that, that representation is accurate so again, on the business lens of things, Boon Ping.

FBP: That quality is important, and securing and protecting are important as well. We talked about some of the data-related challenges across the industry, but very long-term, deep-seated challenges with data quality, and also in culture. And it's not the data itself. It's how the data is treated by  the organisation. We talked about data literacy and  data  accountability as well. How are these challenges  deal with traditional technologies? We’d like to again, get Peter back to get his perspective on how he sees global organisation around different regions, dealing with other differences because of regulatory or market structure differences, that he sees differences in their treatment. If Peter, we can get your perspective on this data-related challenges and how industry players are dealing with them across the regions.

Peter Ku: Yes. The regional differences have the gap  become closer over the years, I'd say,  10, 15 years ago, you would definitely see a disparity between what firms are doing, let's say, in the States or in certain parts of Europe versus in Asia. With the proliferation of the internet and the way we run our business these days and the globalisation of information sharing, conferences, events such as this, we learn from each other. The gap is definitely has become narrower over the years.  What I have seen, however, is that the attitude towards  who owns the problem of fixing data, is still something that I hate to say this, ‘oh, it's someone else in technology that's responsible for fixing the data’. What data governance is trying to do is change the culture of data ownership. Data custodians and data stewardship are  defining specific rules that were really, the value of data governance. When those concepts are introduced, people would say, well, who's a data steward? Like, where do you get a data steward from? Can you hire these folks, and you realise you already have them. They're the day to day business, people that  understand how data is used, where it's used, and why it's important to the organisation. The challenges that the industry faces today are the technologies that organisations have adopted to manage data from database technologies. I was involved with data quality solutions, more than 20 years ago. This is when it was designed for the database developer building data warehouses on SQL server back in the day. Organisations are still using those tools today to deal with  data quality requirements.

For today's business which starts with the business users, they're the ones that run into identifying data errors. They also need to be involved in defining what needs to be done to the data to get it right, and ensure that they're part of the remediation process. They become the de facto data stewards for their critical data elements. The problem is those technologies are outdated, as I mentioned. And it is a risk to leverage outdated technology to help support the people and the processes and the standards of today's financial services industry. The gap is getting narrower as well, in technology adoption. Obviously, with cloud computing these days, they no longer have to buy your hardware and install the technologies on prem. Everything's pretty much on demand. It's really changing the way solutions to manage and govern data are being consumed across the financial services.

FBP: Okay, some of that challenges emanate from technologies that are outdated. So let's hear from Celine from Bank of Singapore. You work with OCBC as well on the AI Lab. On the cloud, you're using all the latest advanced data analytics technology. How are you dealing with all these data challenges? Who owns them, and you as  data innovation officer?

CLC: I agree with what was just mentioned.  I consider the data office and the data management office as a facilitator and the coordinator. We have data.  When defining data, we have data stewards, experts and data custodian. We have all those rules that are clearly defined in the organisation. We have all of those  committees and instances, we have weekly data working groups that are working per stream on a specific project. When there is an issue on the data,  whatever it is,  where we need some alignment in  definition or impact on one data that would impact to derive data for  risk function, for example or finance function then,  we got those specific data working group that is created -- time to align everyone on those specifics but we don't need to include like the whole community. If it's a specific use case that is linked with finance and IT, we don't need to have like all the compliance and marketing folks. On the other way around, when we have some discussion about client ownership, supply and data, we have created a client data chapter. So, which is a kind of, another 1.5 line of defense if you will, that is just here to coordinate across our different markets. What is the definition of our client? How do we calculate what a client and  all the metrics that are derived, so retention, cross-sell upsell everything that is starting from the client view. This is a lot of work and remediation from legacy systems and others, so all these have been prioritised. Metrics have been put  so we are able to handle our data remediation and data quality in a more effective manner, being able to trace that back to the impact on the business.

When it comes to working with groups, OCBC, we also have our group data management committee,  and group data working group, so we're doing that at the different levels within the subsea[1]  itself, and then after, along with the group because of something that will impact  groups. OCBC as a whole on the reporting to the regulator, then,  we need to  be involved. But here we are, one of the participants rather than the one leading the meeting  and the resolution, I would say.

FBP: Okay, great, and very often, so you're a participant. Who leads them, is it technology? Who leads them?

CLC: No, so we have a group data office and a group data management office. We have the same kind of setup within the subsea, as well as within the group. So this is  group data management office that is in charge of  aligning all of the subsidiaries. We have the same kind of governance adapted our subsea level for our specific, I would say, maybe more operational issues linked to data quality that are impacting our business. When at the group level, it will be more regarding regulatory reporting, critical reports and everything that is linked to the regulator. I would say that the kind of users are the same, at the end of the day.

FBP: Okay I understand. Thank you Celine.  Adeline, how do you work with your partner banks in trying to deal with overcoming some of the data-related challenges that we have, as you work with them across some of the analytic solutions?

AK: Thanks, thanks Boon Ping.  I mentioned collaboration earlier and partnership with different parties. For this to really work effectively and therefore effective data sharing to really work, I do agree that we require a global framework and some common standards. Common standards cover common rules, common industry practices, involving the definition of exchange of data. And it's something that Visa, we've done quite well over the years. We've experienced a setting kind of global, interoperable standards, so that, when you use your Visa card in Japan, it works. When you use it to be the card in Bahamas, it works as well. So the same approach that we've taken with setting standards in the payment space. We believe that we can adopt a similar approach toward data and data sharing. If we do this correctly, it then can help with cross-sector data collaboration that will  maximise the benefits of data in the data economy. In doing so, then it offers our bank partners to access new datasets that support better decisioning. Eventually, new commercial models. The one thing I'll highlight is, maybe, digressing a little bit from your question is, again, going back to the consumer, nothing can happen without consumer consent and it's something that we didn't mention, although, embedded in mentions of data privacy, data security.

The notion of the fact that data doesn't belong to any single organisation. If you are from a bank, you would think that the data belongs to the bank. Visa would probably think - some parties in Visa in the past might think, the data belongs to Visa because its kind of should have the Visa rules. But ultimately, especially in the open banking space, the data doesn't belong to a single person, if anything. The consumer has absolute rights to decide whether he or she wants to offer access to the data for better use of personalised services, or developed better products. And that's where consumer consent management becomes important because without capturing consumer consent, there's really very little we can do with consumer data.

FBP: Great, thank you Adeline for that perspective of how you're working with organisations, and also dealing with the more intrinsic challenges of data. Who owns the data and consent? Yong Nieng, you chat a lot earlier about global industry standard for mortgage data. What other data-related challenges are you helping those mortgage institutions with?

TYN: Yeah, Boon Ping, if I can frame your earlier questions slightly differently, in that, how we can free data from the technologies that are tying it down. We will ask my fellow panellists or expert has spoken on technology and the governance aspect. I would like to touch on the role of industry leaderships. Smilar to what Céline and Peter have also commented. It is everyone's responsibility. But in this case, it should be taken up a notch at the industry leadership level. I believe that if the main stakeholders’ industries, and these include the regulators come together and work together to decide what that data models look like- what are the standards and how the data should be exchanged, a very clearly crystallised, the return on investment (ROI), which  we talked about earlier,  in freeing up data, and more importantly, the risk of not doing so. I believe that if this happens, the technology providers and the smart people out there will start to support the industry and develop solutions to address these markets to develop solutions to free up the data. I believe the industry has a strong role to play. They should take a lead in determining how data should be managed.

This also ties back to my earlier comment on data standardisation for the mortgage industry, which at this point, Boon Ping is probably not as well developed, as Adeline has mentioned for Visa and payments. And the benefits of data standardisation for the mortgage industry go beyond just  freeing data and facilitating data exchanges. Because if data can be freely and easily exchanged between organisations, it could potentially reduce information asymmetry, and allow various organisations to make better decisions,  be it in pricing, or in appraising and measuring risk.

Boon Ping, if I can just, tension off a little bit. I believe for the mortgage industry in Malaysia specifically, we shouldn't just adopt. I know there are many mortgage standards out there that have already been developed. Yes, we should adopt, not to reinvent the wheel, but we should further extend it to support Shariah-compliant mortgages. Malaysia is well-positioned to do this,  the Islamic finance is well developed here in Malaysia being one of the largest Sukuk issuers in the world.  I truly believe that industry leadership is the first step in freeing data and decoupling data from the technology.

FBP: As the business evolved. You mentioned the example of Shariah. The other area that is emerging is this whole area of sustainability finance, for example.  So you have green mortgages, as well  Is there a set of standards around that? How does that build around the data itself? We're just kind of addressing those emerging issues, there will be ongoing discussion and debate around that. Meanwhile, let's hear from Robert first and then Monica, on how you see your institutions, or from your perspective how data-related challenges are emanating from the use of maybe traditional technologies. Robert first,

RW: The problems that we see or have seen with legacy technologies is that in data terms, you get a lot of scarcity. You get issues from legacy migrations that have maybe not gone so well, you get all sorts of things happening in the data.

I agree in a lot of ways on the need for community. I have certainly seen in banks that there is a bit of a tendency to work in silos. In the pursuit of standards, we need to maybe standardise our own approaches, standardise the way that we come together on it, and break some of those silos. And that means a little bit of taking ownership, as well for some of those issues to help develop standards, and standards in particular are an issue. That's where I would focus a lot of attention on the people’s side of it and try to agree on ways of handling the things that you see because a lot of the solutions  to some of the problems  take a lot of investment. And yet we still need continual investment in the things we want to do. I also agree with Peter  that the things that the technologies that are learned, lineage is a kind of almost an underestimated technology. It can help in so many ways of understanding. Where did your data come from? If I make a change to an operational system, where is it going to impact? These types of things are quite important. Data teams did have functions and themselves, we'd love to just apply all of those technologies in the way that they know how to solve data management problems. But the prioritisation question about how you handle all of these things comes from the community. 

FBP: Community and internal collaboration, right?   Thank you Robert, and Monica.

MS: Yes. Again, a different dimension. We are focusing more on traditional technologies, but I would like to talk about traditional processes. Here, what about RCDD[2]  process? Traditionally, 15 years before just name, date of birth and address could have been considered more than sufficient to know a customer. The customer due diligence process was so simple and straightforward. But today, the regulations have changed the organisations have big responsibility to combat financial crime. Therefore, this entire customer due diligence (CDD) process itself has become more complex for all good reasons. The data that was considered enough at one point in time suddenly seems insufficient. So this is only getting more complicated in a competitive market with new payment methods, like digital wallets provided by payment service providers. It is absolutely not possible to perform CDD for each and every client and periodically review them in a traditional way. But what we can do is to harness the power of data, utilise data analytics and draw insights and then likewise identify AML risks . So, if you have some other way in which you can similarly detect and manage AML risk, then the risk is taken cared and you do not need to necessarily follow the traditional approach, so it is not just for financial crime. Because we are starting to use analytics big time to detect and prevent frauds as well. With the uncertainty and disruption caused by COVID, this also created opportunities for fraud. As  business processes could be more exposed or people become more vulnerable. Fraudsters want to capitalise on stress and disruptions. So many types of frauds --COVID medicine fraud, COVID loan fraud, misuse of schemes etc. So here, if you look at it, the analytics and AI can be powerful to detect patterns and anomalies and help identify frauds. With evolving payment frauds, behavioural profiling of client and transaction data can help detect fraud. So what is it that is required to overcome data-related challenges? First, quickly look at how much data are we gathering. The International Data Corporation says that the amount of data created over the next three years will be more than the past 30 years combined. Isn't it interesting?  The  World Economic Forum also says that by 2025, it is estimated that 463 exabytes of data will be created daily. So for a layman, it is equivalent to 212 million  DVDs a day. It is also said that 2.5 quintillion bytes of data are created everyday, and it is going to grow. So for any organisation, this creates an opportunity to gain insights and make business decisions backed by facts. The one who has the ability to collect, collate, interpret and make meaning out of the data clearly has a competitive advantage in today's business environment. At the same time, there is a potential risk of data hoarding. If you're not classifying the data appropriately, you will soon be overwhelmed with the data not knowing which is the required data and which one is superfluous. It also brings in additional complexities such as storage costs, the inability to access the right data at the right time, etc. So to me, the data in any organisation does not come all at once, and therefore assessing data quality must also be ongoing. This needs a cultural change. Everybody has spoken about it. It has to be everybody's responsibility with data being recognised as a company-wide strategic priority and we must embed data by design. It is not going to be an easy task at all, but with continued training, awareness and perseverance by all data management professionals. I'm sure we will reach the target state. To summarise, we have three takeaways - classify the data as you receive, monitor data quality on an ongoing basis and embed data by design.

FBP: Okay, great. Thank you, Monica. Doing a different dimension to it not just technology, but  with increasing digitalisation, a lot of data, and processes need to be updated as well.  I mean, new capability, electronic know your customer (eKYC)  in both onboarding. How do you leverage data for those new processes? Now, I'll give you that Monica for your response to the next questions as well. Given the cloud adoption trends across the industry, what are the implication and how financial organisations are adjusting their data management and data governance capability to ensure data continues to be fit for business use? I know you use the other term embed for data design, right?

Leveraging the cloud to access and manage data helps businesses reduce time-to-market

MS: Sure. Thank you Boon Ping, a very relevant question indeed. So to me, the requirement to have good data is agnostic of whether it is inside an application or a data warehouse or a data lake or cloud, etc. So we must make sure that the traceability of data is established back to where the data originates and gets processed, irrespective of where it sits. Cloud adoption definitely means it is going to help us big time because AI and ML are going to get access to big data, and they will be able to perform a threat analysis that's giving us more insights to data. But there are many other surrounding risks that we must address in parallel. Because I want to keep it aside on data quality, per se, because there's quite a lot of other risks that could pop up because of this, such as data sovereignty risks, data protection, security risk, resilience, etc, as an example. So, especially when it comes to cloud data sovereignty, its definitely a cause of concern. We must watch out and make sure that we have the right balance of controls that are implemented because data sovereignty is originating because of the geopolitical situation. You also have the responsibility to make sure you protect personal data, and obviously, security becomes a cause of concern.  So I will just stop it there and give the opportunity for others to add on. But this is definitely an exciting topic.

FBP: Okay, great, thank you, Monica. Next, we'll hear from Celine on this issue or this opportunity that is afforded by cloud adoption.

CLC: Cloud adoption is not only an infrastructure topic. It comes to business.  Cloud is bringing velocity, is bringing,  AI as a service through API. I mean, like training, a text recognition or voice recognition algorithm, takes a lot of data. Some people have done this for us, whether it's Google, AWS, Azure, or the other players. At the end of the day, we need to be able to define what is our core business, and what is,  kind of too complex for us to  be able to build and manage internally. Cloud is an enabler for faster business, insight,  deployment of AI through,  ML ops, and then,  automated machine learning platform, so that enable faster adoption of analytics and insight. There are some challenges as well. We just talked about data sovereignty because cloud like any new technology requires an upskilling,  and a change in the process. I mean, there is a question of role and responsibility that the cloud framework is bringing, as well. When it comes to who is in charge of the security of the data of the protection of the data, and it's a kind of a new mindset, where all of our policies  need to be adapted, our internal processes, because, it allows so much more flexibility than what we had in the past with our on prem platform, but also, a certain level of confidence from the regulator, and this is where, again, I will mention MAS, was one of the first regulators to implement a cloud adoption framework -- telling that it's okay to use Cloud to the financial institution. But even with that,  there's still a  mindset that things are more protected when they're on your own data centre. But, when you look at the capability of a cloud provider like the level of cybersecurity, they have, like 100 hackers that are hired full time 50 of them are hacking the cloud, 50 of them are protecting their cloud, who in the financial services industry can have enough resources to be able to implement such security?

Like a cybersecurity team of professional hackers.I mean, all of those new resources and skill sets are very rare and seldom and so we'll never do better than when people that are specialised to do are doing so it's more of a question of mindset,  and being comfortable and adapting our processes, policy, and then security guidelines, according to new technology.

FBP: Okay, adopting the mindset. That's alright. You mentioned the opportunity that technology affords.

RW: Singapore  was faster off the market in getting cloud established as a means of doing business and Malaysia is just playing catch up a little bit. From what I've seen, there is a little bit of attraction there. But the issue is around the confidence level and not that it's not the right thing to do, it’s clearly is the right thing to do. Because you get to  where things are going. But handling it governing it, there's still a little bit of way to go to build that confidence level in taking that there's  what is quite a big step.

When we started this meeting, I was saying that the there is a gap between having access to advanced technologies and the business’ ability to exploit those technologies. This is probably a good example of it because even the technologists themselves have got a little bit of learning to do and making it right and safe and unable to be utilised in a big way, but certainly, the right thing to do to make that big step. But  that's where I feel that Malaysia is at the moment, we're starting the journey.

FBP: Great. Thank you, Robert. And finally, Peter  in cloud adoption, some of the risks that Monica mentioned, data sovereignty, data residents, sea of data centre, and maybe for cloud provider that there's only that few, and there's also a certain concentration risk of cloud provider.

PK: Yeah, there's a recent industry study that measured that 87% of new compute workloads across financial services is happening in the cloud. There's not just one cloud. It's not just Google, Microsoft and Azure. I mean, excuse me, AWS, it's not just one application vendor, it's potentially many. And the reality is more and more of the infrastructure platforms and applications along with analytics, those net new investments are in the cloud these days. The challenge, however, in financial services is that a lot of the systems that they were supposed to replace that ran on premise, they don't go away overnight. So what ends up happening is you add layers of new technology on top of legacy technology which complicates the data problem. So as more organisations are modernising their business capabilities to the cloud, there are some things that they have to keep an eye on.

Identifying clear objectives to utilise and classify data to align with the organisation’s strategies and ensure a customer-centric approach 

Number one, how am I going to get all that data from all the other systems into the cloud investments that they're making? How do they ensure that the data is going to be available, where they need it most? Number two, how does it impact governance and accountability? In data controls, as business application software as a service solutions, IT organisations used to go crazy when the business gurus would say, ‘Hey, by the way, we just signed up a multi-year deal with XYZ vendor’, and all of a sudden, we now have a new finance application and we need it all wired into the rest of the business. Who's going to be accountable for that, who's gonna be the data steward across both the data as well as the application layer. These are things that organisations are going to deal with. They're gonna have to operationalise data governance, so they can really control, monitor and adjust the governance and accountability requirements, cataloging and classification.

We talked about lineage, Robert, it's hard enough to get lineage views on  premise systems. Imagine now as you have this hybrid world. It's a matter of being able to access all that metadata and being able to  trace,  where those dots are connected, accessibility and usage, entitlements and access. We talked about data consumption. Who's consuming data from the systems and applications in the cloud that has to be monitored, from a privacy standpoint, from a data security standpoint and  also, from a cost standpoint, because many of these cloud providers charge based on consumption. The more you use, the more it's going to cost you money. So over time,  is it really cheaper to go to the cloud? It all depends on what you're using it for. And then,  privacy and protection.

I'd say 10 years ago, I remember sitting in a room with the major banks in New York City. They treated cloud computing as sort of a science experiment. Hey, we don't really trust having your client data in the cloud. Let's fast forward to today. No one's having those conversations. Jamie Dimon just the other week talked about how JP Morgan is spending $12 billion on technology. They have over 500,000 technologists employed by that company more than AWS and Microsoft combined. And so you have to think about, the amount of data that they're creating and consuming. Finally,   that architecture of integration, data quality management, data lifecycle management, these are all things by the way that Informatica has been focused on. For the last six years of our company, we have transitioned our own technologies, from the traditional on premise solutions to all things clouds. Everything I talked about earlier, enables us to address the requirements out there. These requirements are going to become more complicated. It's not going to get any easier, especially for regions like Malaysia, as you mentioned, Robert, as you guys are now building out your cloud ecosystem. And as you know, the challenge is making sure that some of those ecosystem providers have physical operations within country because those are the requirements, and you're gonna see more and more adoption over time. The message here is, don't neglect what you need to do with your data, because it's gonna get harder to do as more and more companies invest in the cloud.

FBP: Okay, great. Thank you, Peter. It's been a very thorough and intensive discussion that we have this morning on how financial organisation like yourself, building into data businesses, identifying and recognising the value of data as an important business assets and how to put that together requires collaboration. That data is not just the domain of technology of data scientists and on getting the data right, and using the data requires a whole of organisation effort.

Collaboration, we keep getting back to that word. It requires technology, it requires governance and it requires our business users to understand and to own the data and the processes. And also creating the structure, the talent, the skill set to manage that. It is an ongoing challenge. Technology is there to enable API, AI, machine learning. Cloud offers new opportunity and capability to address some of the challenges that we mentioned. To Robert's point that a business must be able to identify what is their objective of becoming data and customer-centric. How is that aligned to their strategy? How does their daily data management governance effort go to achieving that strategy to be able to satisfy their own requirements for return on the investments that they put into data and technology. And we mentioned that it’s a continuing and a growing investment and one that the business needs to see return on. There are many complex issues involved that is ongoing. The enabling technologies that can help meet some of those. So this is not the start, this is a continuation of that journey and that conversation.

At this point, I want to thank Celine, Monica Adeline, Peter, Robert and Yong Nien for sharing your insights and experiences. We hope that our audience, whether you joined us via Zoom, Facebook or LinkedIn, you found this session insightful and useful. I'm sure we have the contact of our panellists, if you'd like to follow up on their insights.

If you miss any part of the session, or would like to have a playback you can visit The Asian Banker RadioFinance website, where you can download the recording of this session, as well as other sessions in the past that you might have missed. And to find out more about the future sessions that we will have. So until the next event, we wish you all a good day. Thank you for joining us.