Reimagining Data Functions in the AI Era: Fixing Ownership, Governance, and Trust

FULL TRANSCRIPT

This transcript is based on the episode’s English auto-captions and has been formatted for readability. Please allow for occasional transcription errors in names, acronyms and specialised terms.

[00:00:00]

It's now more important than ever before to be conscious uh rather than you know to be making conscious rather than unconscious decisions about what data you're collecting, where it is being stored and shared, who has access to the data, who owns the data, what are the risks and more importantly what are the controls around managing those risks. So the real challenge facing data leaders now is >> [music] >> uh how to build an effective data culture uh not just within the data and technology teams but also across the rest of the organization and [music] that's a tricky thing. So if you feel that you own the uh the governance, you own the data security, you [music] you are accountable for that, then you will make potentially you'll make different decisions or or you'll make better decisions for [music] the organization because you will not just think about the business outcomes in terms of financial outcomes or sales etc. You will also [music] think in terms of the risks and you know the threats that come along with that. And so I I think the traditional models around uh fixed you know ongoing structures uh I think will start to crumble as we move forward. uh it'll be become more and more uh flatter structures where teams are more and more getting you know the autonomy and empowerment to self-organize the people are feeling the pressure to engage with AI and move forward [music] and show some runs on the board and deliver some outcomes and be able to talk about AI outcomes etc. And while that's a good thing, I think you got to [music] move and certainly start thinking in in that direction, but you cannot ignore uh you know the safeguards [music] and and the risks.

[00:01:55]

Hello and welcome to the enterprise tech talk podcast. I am your host Saumitra Kalikar. Now the topic of today's conversation is something that is firmly on the agenda of CIOS, cos and increasingly on the board agenda as well and that is how to moralize the data conscious to deliver uh strategic businessups and this is because over the last decade or so organizations have invested in data capabilities but has struggled to map those investments to quantifiable business benefits. So the so the question is what fundamentally needs to change within data functions to re to deliver real business value and to help me unpack this important topic I'm joined today by Surit Sethi. Surit is a senior technology executive uh with deep experience in uh establishing and and managing data functions in large complex and regulatory environments. Sit welcome to the broadcast.

[00:02:53]

>> Thanks. Thanks thanks for having me. Many listeners would know Surit and I have worked in the past. Uh so Sur it's been a while but uh what's happening? How how are you now this these days? >> Yeah, good. Yeah, it's it's been about 6 months hasn't it Saumitra since uh since we worked together last. Uh I've had a bit of a break in that period. Um and I've been really uh enjoying the time to refresh, recharge, travel and also to reflect a little bit on what I want to do in the next phase of my life. So it's been a really good opportunity to take a step back uh from the day-to-day and just figure out what is it that uh you know uh that'll give me joy for the next period of my life. uh before we get started sit um if if you don't mind if you can provide a brief overview about your background or areas uh particularly in the context of today's topic um and maybe you can also tell us how you got involved in the in the data domain what excites you to be in this domain >> yeah thanks so uh I guess a little bit of an introduction first so I uh I've been working for like I have experience over 30 years uh and it sounds like a lot and you feel like an old man when you say something like that but time has gone really quick uh and I've had the pleasure of working across different geographies so I worked in in India first uh then I worked in Australia uh I have I started my career in the business because my background was a business background I did my bachelor's in commerce and my MBA in marketing and finance and then uh I joined uh A&Z uh to start my career uh in banking And uh that was on the business side.

[00:04:37]

And then over a period of time I I kind of you know moved into technology uh maybe by accident right because uh I was secounded as a as a businessme onto an IT project and uh and I really enjoyed you know helping technology to understand what the business requirements uh would be. That role was a business analyst role to start with and then uh um I just found that I had that unique ability uh you know given my business background to actually interpret the requirements uh in a in a way that the technology teams would understand. So uh and I had some really good feedback. I had a good experience. So I kind of stayed in technology uh from there right. So I moved into a lot of technology management roles since then and I've worked across different industries uh in financial services as well as uh health insurance where we last worked together as well. Um my last role was uh at Medbank which was general manager for data. Um and that was a fun experience. Uh uh I can say now looking reflecting back uh but a great learning experience as well because I moved into the role uh just a couple of months before the uh cyber incident happened and uh and that was quite a roller coaster ride as you know uh we went through that together. Um but the learnings that we've had since then and uh the achievements that we've had you know as as a as a collective as a team I feel like I'm now ready to uh to share that knowledge and uh you know help other organizations as well. So that's that's part of my plans uh for the foreseeable future as well. uh and to and to help that uh you know help me prepare for for that I have actually been uh investing my time over the last 3 months doing my GICD uh you know qualification company director's uh course um I've appeared for the exam I've done my uh assignment as well and I'm just waiting for the results so uh fingers crossed over there that uh you know I get that over the line and then I can uh get on to uh you know potentially getting on to the boards of one or two companies equally uh doing some consulting and helping companies to uh you know to share the experiences that I've had and uh and help them to uh you know to manage their data better. Let's start with a uh a broader context with uh so so if you look at the data function right uh or the how it has evolved um historically I remember data function was primarily considered as a reporting function within the organizations it used to to deliver responsible were delivering reports to all parts of business operations right um but as organizations started focusing more on customer experience to personalization And besides growth in AI, that means um data has become a really important strategic asset now right and uh what it means that the expectations from um CIO, CEOs and even board has shifted um or office.

[00:07:47]

You have been in this industry for a long time. So I wanted to get your perspective as to how the expectation from a role your role like a head of data function role has evolved over time. Yeah, that's a great question Sumitra. It's something that I have had the opportunity to reflect on as well and uh while I was doing my GICD course as well I you know I through the simulation exercises through the assignment etc I got the opportunity to put myself in the shoes of of the board and the directors and how they kind of view this as well. So uh so what I feel is that it's it's you know the expectations internally are a function of how the external environment is changing around us right so so the the internal expectations are influenced now over over the recent past u by the changing expectations from both customers as well as regulators.

[00:08:48]

So people are now more aware than ever before. uh you know not only people like you and me who are in the data and technology domain but also you know uh the common man u and women just just on on the street everyone is more aware of the need for privacy and for data protection and they are expecting more from organizations right so the cyber incident uh you know at my last organization occurred u as uh as we spoke about you know in a short period after I moved into the role but the experience taught me about managing expectations uh not just from customers and regulators but also from the staff involved because these things have a big impact on staff and the staff feel a sense of responsibility around managing data etc. Right? So data leaders now really have a balancing role. It's a fine balancing role where they need to manage expectations of internal stakeholders but but it has to be in the context of a rapidly evolving external environment relating to cyber threats uh and the opportunities and threats that come with AI as as uh as you're aware while also balancing the needs of uh the people that you're leading and and you know you have a you have a large team typically when you're in these sort of roles and the people are looking at you to guide them and to support them in a constantly changing environment. So uh so it's a multifaceted uh and therefore very exciting role to be in uh these days. Another role that has I think evolved more recently for data leaders uh is about educating their peers and the business teams as well about how to manage data effectively and what data ownership really means because we talk about data ownership and the fact that you know uh there has to be a clear owner of data. business needs to own data but but the concept is not very well understood uh and not consistently understood across business areas. So I think as a data leader that's that's part of our responsibility to get that alignment and understanding across the organization on what it really means for for them >> and um you you briefly mentioned this but this is I wanted to unpack the fund more as to it looks there is a a healthy tension emerging between when it comes to the data asex uh or how you manage the data of organization because at one side or there is a emerging focus and acknowledgement that data is strategic as um and how how we can use it for better customer experience and innovation and competitive advantage etc. At the other side you as you briefly mentioned um data also possess significant risk in today's world um particularly when it comes to managing customer customers privacy um um operating within customer allot concepts operating within your own uh risk appetite and uh some of these things and responding on to regulators.

[00:11:38]

Uh and again as a as a um head of data function um the how do you see organizations should balance that out uh enabling the innovation etc but at at the same time make doing it safely >> yeah that's an interesting one isn't it uh look uh I'm sure everyone's heard about data as a strategic asset there's been a lot of conversation around that in in you know in recent times um I've heard people say data needs to be treated as oil uh it's probably even more relevant with what's happening around us in the world at the moment and how expensive oil is becoming and how precious that is. So it's probably not a bad analogy uh you know to think about but what sometimes people don't realize is and focus enough on as you're saying is the data can also become a liability if it is not managed properly. So uh with cyber incidents, privacy laws, customer expectations, rising costs of managing security etc. all become relevant. And therefore, it's now more important than ever before to be conscious uh rather than, you know, to be making conscious rather than unconscious decisions about what data you're collecting, where it is being stored and shared, who has access to the data, who owns the data, what are the risks, and more importantly, what are the controls around managing those risks and managing the data in place? And and you got to look at this strategically and not just every person doing it for their own business area. It has to be up to the executives, up to the data leaders, but also up to the business leaders and the board even to uh to be strategic about uh you know what risks are acceptable, what are unacceptable, what is the risk appetite that the that the company has and then to manage uh you know the data collection, the data storage management and security and controls in that light so that everything is being done as a conscious decision rather than just happening around the company, you you know two different levels of control uh depending on how the respective business leaders look at it. So it has to be it has to be driven at an enterprise level and it has to be driven at at very senior levels and communicated very clearly across the organization um which otherwise it can very fast become a liability.

[00:13:48]

>> Yeah. Yeah. So, so uh on that point, let us discuss the the concept of modern data function. Um building on what you just said um because uh many people see that modern data function means investing into new technologies which is of course part of part of the consideration but that's not everything in my view. uh you need to look at u the operating model aspects you need to look at uh uh the skills aspects and many other things right uh relationship with the business etc so uh do would you mind painting the highle picture about when we talk about modern data function >> yeah that's a good question as well so uh look as as I think you were alluding to traditionally modern in technology world uh you know uh was spoken about in reference to how how new the technology was whether we've refreshed the technology whether you know uh it's been upgraded uh whether you've got u you know support in place and things like that but I think uh as we've been discussing modern technology functions and modern data functions specifically uh are expected to be more resilient and managed proactively as well so there is the management aspect which is becoming far more important uh you know the the whole framework around that rather than just the technology itself. Um and therefore it's far more than just a technology refresh. Now it includes all aspects around uh clear data ownership um risk management framework all those things being in place. Um I'll also add that a real challenge facing data leaders now is uh how to build an effective data culture uh not just within the data and technology teams but also across the rest of the organization. And that's a tricky thing, right? It's it's not because culture is very diff very difficult to define and very difficult to measure and especially across large organizations.

[00:15:43]

Um so everyone in the organization for example has to be aware of the consequences of collecting, storing, sharing data. So in that in that sense there has to be a real consciousness around what is the data culture in the organization right? How do people view data? how how do they view awareness or their responsibilities around the data? Because I I firmly believe that managing data is everyone's business. It's not just you know the responsibility of of a small data governance team sitting in one corner. I had a small team. At one point it was three people and then it grew from there. But the the key thing for uh for me uh in that role but equally for other data leaders uh you know uh in in other organizations is is around how you harness that energy and you know uh get that consistent understanding and awareness and build that level of knowledge and education around the organization so people understand what it means they know where to go for guidance and and they know exactly uh you know what what they are allowed to do what the boundaries are uh because you know especially with AI coming in it's important to have principles uh policies around and and a and a control framework around how you're using data. So that that's I think part of the whole modern technology or data function now because it can't exist with just technology alone. It has to it has to exist coexist with a with a proper framework around it which has risks and controls education culture everything has to be a part of that and and I think one one thing which which is still evolving in the industry is around uh consequence management right so so people people will will say and do and try and do the right thing by and large around data however unless there's a real consequence management uh framework that's built into people's responsibilities and that there are consequences around uh not complying or uh you know not being as resilient as you want people to be depending on the culture and the risk appetite of the organization then um it's just not going to get taken that seriously as as it needs to be around uh you know potential misuse of data.

[00:17:59]

>> Yeah. Let's um uh speak to focus a little bit on the the operating model uh about data function as well. because that is something I wanted to pick up because um I mean uh all the IT functions right uh historically uh not it's not only data function where you take any IT function historically have largely been very central functions right uh your central data team central architecture team centralized separate teams etc they would deliver services to rest of the officers that how the certific operating model was but the broader IT operating models are now emerging are changing. There are new patterns emerging where you're talking about more federated operating models where part of the IT functions are being embedded into uh other parts of the business to allow them more autonomy etc. Right? How do you see data function from operating that operating model perspective evolving uh what patterns are emerging and uh maybe pros and cons of patterns as well?

[00:19:03]

Yeah, it's it's certainly an area that's been evolving, Sumitra, as uh as you've seen as well. Uh look, I think uh what we are starting to see increasingly work well is is the emergence of uh empowered end-to-end teams, you know, within an operating model. So traditionally we've had uh a clear distinction and separation between baseless and IT teams and data functions and and and the data warehouse team sits separately from the data analytics team and the reporting teams etc. So now I think increasingly what we are seeing is that uh uh end to-end teams that are empowered to deliver business outcomes uh tend to achieve more and that model tends to work better than the traditional model where there are a lot of handoffs. Um it's it's but it's important to have uh have awareness around the outcome that and clarity for the teams around the outcome that is being delivered there. Otherwise uh you know traditionally what's what's happened is people have become um almost slaves to uh you know to uh to the funding uh question to uh to how we structured and the constraints that we've had and people try to do the best that they can within the constraints that they have and that's how they've been measured right uh but but I think you get the best outcomes and increasingly we seeing organizations tend towards this model where uh you build teams around specific outcome come. And the teams have to be therefore uh agile not just in terms of how they're working day-to-day but also in terms of how they're structured because once you deliver one outcome then you then you might need to reorganize yourselves uh you know uh to to kind of uh structure around the next outcome and and how to deliver that in the best possible manner. So, so I I think the traditional models around uh fixed you know ongoing structures uh I think will start to crumble as we move forward. uh it'll be become more and more uh flatter structures where teams are more and more getting you know the autonomy and empowerment to self-organize around uh around outcomes and uh and then they measured on the outcomes that they're delivering not just around uh you know working within specific constraints and uh how far they moved etc. It's it's either you're achieving the outcome or you're not achieving the outcome. And and therefore the uh and one of the things that has to evolve to kind of support that model which which is uh which is a shift from the traditional models is around how teams are measured and how they're rewarded as well. Right?

[00:21:46]

So uh you know how targets are set and how people are uh are encouraged to compete almost with each other. Right? Because you have these ranking systems in organizations where people are compared against each other. So if you have to do better than someone else then someone else has to do worse than you by definition. So I think those models have to uh and HR functions are increasingly becoming aware of that and uh and working towards that towards resolving that because uh you got to encourage business outcomes uh to become the focus rather than individual performances. So uh but that that's a challenge that I think we got to work closely with as as leaders and as data leaders. We got to work closely with our HR partners and uh and solve that conundrum as well. >> And in that future um do you see there will still be some some centralized data team in some shape or form um and if that is the case what could be the that centralized accountability um and then what roles and what accountability might be fed out >> I think uh it's it's not going to just all disappear. It's uh there there's while we have truly end to end teams delivering there is um I think the question of what sits in the business and what sits in technology etc becomes less relevant. However, regardless of where it's sitting, you still need to drive uh you know uh a certain structure around you know what what is the data strategy, what is the governance framework within the teams are supposed to operate, what is the architecture and how is that going to be taken forward, you know what's the strategy around that, what are the controls that we're implementing, what are the policies and you know processes that are be put into place. I think all of those functions uh has have to still have some central some sort of central ownership because you can't you can't just let every team decide this for yourself. You got to have you know uh an organizational data strategy. You got to have a governance framework policies controls etc which align with the organization's risk appetite with the strategy of the board and the strategic outcomes that the company's striving for. So I think a lot of those functions monitoring auditing that putting governance around it uh you know even though you you want to have an an element of self-governance but there has to be some amount of you know second level third level uh functions that are auditing governing uh the use of data and how these teams are working and and not and that's not just from a risk and security perspective. also from a consistency from a continuous improvement perspective so that uh you know we're all moving forward at the same pace and uh and consistently as an organization.

[00:24:24]

So some of those functions will still uh need to be driven centrally by >> uh let's move on and um f uh discuss a little bit about the economics um uh aspects of how you managed investments within data function. Um now historic historically like many other IT functions most of the init functions are very project based and in to be honest in most of the organizations that still is the case but as you see some of the more mature organizations where the data functions are more mature they started embarking the the the new approaches of building data products for example and they're having more product based investment approach not necessary is a project based do you think most of organizations will eventually pull that way interest defined will define data products and product investments and funding products instead of funding in data projects to deliver into categories >> look that's an that's an interesting one as well because um I think as you rightly said uh that traditional funding model has been you know has been getting challenged uh over the last uh two three years uh increasingly so and we've seen We're seeing the you know similar to what I was speaking about the HR function kind of evolving and aligning with how how things are moving. Uh we're also seeing the finance function similarly aligning and uh you know uh adapting to the new way of working and how uh how best they can support uh you know the business and technology and data teams to achieve their outcome. So I think that there's still a role for project funding in some cases. So uh it's it's uh you know as an example where you're where you're replacing a major platform or you're building a completely new cap capability that's across the enterprise uh and especially if if there is a strategic investment involved. I think that there's probably still a role for uh you know whether you call it project or uh an initiative based funding purely from the perspective of uh being able to monitor the and you know the achievement of business benefits and and make sure that we're making the right decisions and and controlling some of that spend.

[00:26:38]

However, more and more uh as I think you were referring to as well, the the funding model seems to be evolving towards uh towards product funding. However, it's it's it's not just simply product funding. I I would I would call it it's more around funding the teams that we were speaking about. So if you if you speak about building end to-end teams, so I think uh what we seem to be evolving towards is having funded teams in place uh which are driving around you know driving business outcomes based on business priorities. uh and once those teams are funded then I think in terms of driving the right behaviors in the organization it helps to drive the right behaviors because then the conversation is around what's the best priority and what's the highest priority for driving the best business outcome rather than where's the next thousand or 10,000 or $100,000 funding going to come from right because we've seen in the past a lot of those conversations can become involved it can take months it can be unproductive where people are generally you know basically organizations doing business with themselves and trying to find out where the funding is going to come from how are we going to charge them out um I think we're moving away from that which is really good to see uh um so in in that sense if we have funded teams whether they are funded product teams uh as you're referring to or uh you know they are structured around driving another business outcome uh I think that's that's probably going to be more and more uh what we will see going forward uh unless there's a major you know strategic initiative that is being implemented uh which I think by the nature of uh the amount of investment that the board is making they would probably like to see some amount of control and reporting back on how much spend is being uh you know approved to uh to deliver what outcome and whether the business benefits are being realized or not. So I think it probably going to be a mix but we'll see less and less of project funding as we move forward. Yeah. And u the the question is uh about measuring the value of of these investments. So what are the key um uh OKRs or KPIs that you think will still be relevant going forward? Um when it comes to data investments uh with are those around me around risk compliance or or more about growth? Yeah, I think that's uh that's a very uh organization specific uh question which can have different answers at different points in time. It's also it's also a function of where the organization is at and what the what the priorities are that the board has uh you know has has decided uh that the organization needs to focus on.

[00:29:32]

So um I think it's it's every company has to decide that for themselves but uh the key thing here is that these decisions should not be based on what's been done in the past and you just keep doing the same. It has to be consciously you got to consciously sit down um and increasingly it's important to be nimble and agile around the changing environment and being able to adapt to that. So I'd probably say once a quarter you need to sit down rather than once a year as we've done in the past uh to determine what what the next uh you know quarter's uh targets should be, what the priorities are, what the board is expecting, how the external environment has evolved, what have we achieved and what have we learned from the last quarter and how do we actually want to focus our energies and uh and our funding and the people on the next quarter and then uh you know continues to keep evolving and adapting to that because the answer in the first quarter might be different to the answer in the second quarter depending on where you are in you know in in your journey uh as an organization.

[00:30:30]

>> Let this let's shift to our focus on the next um uh area which I wanted to explore with you that is around the data governance right and uh um of course nowadays uh trust has become a very important uh consideration at the board level as well customer building customer trust and data governance data in particular plays very significant role there. Um uh but at the same time um as organizations want to um be more agile and more innovative and experimentative, how do you see the data governance as a function evolving um where they need to be accountable for certain standards that etc safety of the data but at a rate of allowing this to be more agile. >> Yeah. So I I think uh the way we look at uh traditionally you know I I think if you had asked that question maybe a year two years back uh the answer would have been more around what the data governance team should be doing and how they should be driving the governance and uh outcomes. uh I think increasingly now as we've been discussing the uh uh the the expectation is more and more that governance is is built into how we do things rather than a separate function which you know uh which is responsible for that. So everyone is accountable for building governance into the way they operate into the delivery functions uh rather than just having it as a separate bait down the track. So it's I I feel it's it's similar to how software testing was viewed a few years back right we used to have development completed and then moving over in a waterfall approach into testing and then you know we used to move forward but then increasingly we realized that the later you found bugs and the later you found issues and defects the the more costly the more difficult it was to uh to fix them right so I think governance is going at our going in in a in a similar kind of fashion now where We're finding that it's it's too late uh you know as an after afterthought to uh you know once processes are completed delivery models are are built to then put governance on top of that we have to think about governance as as embedded and as part of the way it's just how we do things right so uh so every process every business function every uh data function has to have an element of governance and element of uh control built within within the function itself rather than um separately because the expectation from the regulators as well as our customers is is nothing less than that. They they want to see not just how we are responding to incidents and events but what are we doing to actually prevent them from occurring in the first place and that can only happen when governance becomes a way of doing things. it becomes embedded into the culture itself into our DNA and we think of risk and we think of governance when we're thinking of driving a new process or a new product or a new upd and it's just a part of how we do business >> and the whole new governance model or you see the ownership of business data customer data residing uh um and do you think that's that's the right approach to basically make business accountable for the data they are more responsible for you they they use uh while uh uh applying the the common governance standards that data governance team will be establishing.

[00:34:03]

>> Yeah, I think it has to be it has to be Sumitra. Uh you know uh I'm sure you'll agree that uh you know when when you see a separation between the collection and use of data from the responsibilities around ownership and governance then it just doesn't work. you know, it's uh it's it's like you can you can have one small team somewhere accountable for driving a certain outcome um um and not having any any authority to control uh what's happening up front around uh around around that function and it's never going to work. So it's it's the same with with data governance. It's the same around security. Unless the business leaders, unless the business teams are educated and are aware and and and and the ownership resides within that business domain or the business area which is actually making the decisions on what to do. Uh it's no different to what we were just discussing on the last question, right?

[00:35:04]

It's you got to build governance into you know your processes and what you're doing and and that that'll only come with ownership, right? So if you feel that you own the uh the governance, you own the data security, you you are accountable for that, then you will make potentially you'll make different decisions or or you'll make better decisions for the organization because you will not just think about the business outcomes in terms of financial outcomes or sales etc. You will also think in terms of the risks and you know the threats that come along with that and you'll make more balanced decisions. So I I think it's important to have that ownership reside within the business where it drives the right thinking. It drives the right behaviors. It drives the right outcomes and and you know it just becomes a way of governance and just becomes a way of thinking and it helps to drive the right culture. All of those things that we've been discussing and that can't happen if uh you know if there's separation between who's actually driving the function versus who's accountable for governance around that. It has to go hand in hand.

[00:36:08]

>> Yeah. Yeah, we we discussed about your culture and the other important um uh point I wanted to unpack with you was about uh the the skill the talent within the data data function uh um uh there in in the in the historically the data function had different different skills right from data architecture, data engineers, data operations, data stewardship and so on and so forth right um um as in in the modern data function that we are uh thinking talking about here. Are there any new skills and talent needs emerging? What are the gaps there? Uh important gaps that that you would like to highlight that organizations should focus on. >> Yeah. So Suta the way I look at that is uh is is uh you know in in the context of what we've been discussing uh as well. I think the gap the biggest gap that I'm seeing at the moment which which data teams and data leaders are increasingly required to play you know a leading role but equally other other leaders have to uh you know play a role in that is is around building the data literacy and awareness across the organization. It's not just I I don't think the gap is specifically you know there's a big gap in in the data teams as such but but there is a big gap in data literacy awareness governance culture around the whole organization outside of the data team right so this is where I think it's really important for us to face into that as as uh executives as as leaders uh you know not just in technology and IT but also sort of technology and data but also in the business and and also at the board level where we face into uh you know uh the kind of questions that we've been discussing before around how do we build that whole culture the whole framework the entire consequence management the data literacy awareness so certainly data teams have a have a significant role to play uh even in my last uh role and the last organization my team was playing a a huge role in in educating the business and lifting the whole awareness around uh the use of data and and you know uh the governance controls that need to be uh in place and and and driving the ownership and responsibilities etc. So I think that is we've got to continue the momentum on that. We've started to build that and I'm you know in the uh seminars and conferences I'm attending I'm seeing increasingly data leaders are talking about this and and that's becoming a consistent theme. uh so we've got to support each other, work with uh our business counterparts and and you know continue that momentum and and drive that because that's where I think the real uh benefits to organizations will will be in the future. um specifically around technical skills if you ask I think I think it's probably not not a great surprise to anyone to uh to hear that you know AI is is the one area which uh there's still a little bit of uh uh little bit of confusion in in some instances around okay how much to engage how much not to engage uh what skills do we need uh and it's such a rapidly evolving space that uh you know that's that's where I think you got to sit and consciously make makes our strategic choices around how do we want to engage uh and I was reading just this morning on uh on LinkedIn I think somebody had posted a really good article which uh which spoke about uh and they were saying you cannot have an AI strategy unless you have a data strategy um and I think that's really true uh because you got to you got to work out how you're managing the data and what you want to be using the data for before you can apply the AI lens on top of that right and and And it's important to uh make sure that the data that you have is reliable and the quality of the data and you know uh the reliability of the data and the trust around the data exists before you start driving AI outcomes using that data because if if you're using the data that you could you would trust and no matter how much AI you apply on top of that it's not going to give you the right outcome. So uh so I think AI and data uh strategies and skills and awareness and policies uh all have to go hand in hand and I think that that's an area that's still evolving where there's there's still uh a bit of confusion in organizations.

[00:40:26]

>> Yeah. Um and uh you as you rightly said um most organizations are moving ahead with VTI strategy but not necessarily giving enough attendance focus on the data strategy and uh it's no surprise that um there was a report from MIT or somewhere that um 90% of organizations they they are doing AI experiments and and pirates etc but they're struggling to move those pirates and experiments into production and that's where they are failing. So and um one of the reasons potentially is the readiness both from um use case definition identification perspective but also building the uh the backbone I would say from data and technology perspective right so what's your advice to the those people who are executives who are moving ahead with the AI strategy um as to how they should actually temper their expectations um and uh to make sure that data security and the rest of the technology backbone is is well prepared.

[00:41:30]

>> Yeah. And I think that's that's that's a really important point uh Sumitra because it's it's very tempting and and at times I'm finding that people are you know business leaders increasingly are are feeling the pressure from what's happening out in the market and what's happening uh you know what some of their competitors might be doing or what's happening internationally that people are feeling the pressure to engage with AI and move forward and show some runs on the board and deliver some outcomes and be able to talk about AI outcomes etc. And while that's a good thing, I think you got to move and certainly start thinking in in that direction, but you cannot ignore uh you know the safeguards and and the risks that come safeguards that need to be in place and the risks that come come with this uh at the same time. Um however, it's also very easy and and I'm finding some organizations are are doing the exact opposite as well where they just saying no sorry we're not we're not ready. We're not going to engage with AI. uh it's too risky, it's too new, it's unproven and I'm not going to do it.

[00:42:34]

And I think the answer has to be somewhere in between. You got you got to balance it out. It has to be it has to be customized for the organization. The risk appetite of the organization where the organization is at bit any point in time uh what the strategic outcomes are and the priorities are that that you're focusing. So you got to make the choice and and you know make your own decision but but do it in a strategic way where you have a clear uh target in mind and you're moving towards that and you're doing it as an enterprise and you're communicating very clearly and taking your staff you know uh and your customers on the journey as well. Um but but as you as you were alluding to as well and I think of what we spoke about before to be able to get the best outcome from AI, you got to make sure that you got trustworthy data sitting somewhere and reliable data. So so I would I would say the first step to engaging with AI and driving better AI outcomes would be to to do an assessment of where you're at with your data. Uh what you know what uh what has been the experience of the business. Talk to your business. talk to your customers and and understand what the issues are, whether the data is trusted, where the issues are, what needs to be done to lift that level of trust and governance and security or whatever it is, wherever it is that you are at at that point in time and then build AI layers and AI interfaces on top of that because then you will you know from day one you'll have reliable outcomes using AI but you will also have safeguards in place uh where you know you your your risk framework works your security frameworks are all robust enough and you got policies in place you got strategy in place before you start you know playing with those shiny objects but that you know it's very tempting otherwise to do >> okay so as we wrap up sit I wanted to get your your advice on a couple of important points >> u so if uh let's say somebody is uh newly being appointed to uh head of data functions in the current uh environment where uh which there is a lot of focus on AI except and there is a lot of focus on modernizing data functions. What will be your your guidance to those people who are just getting into into the store? >> Yeah. So I I I think the advice is is not very different to what it would be to any new leader, right? So whether it's in the data space or you know you're moving into another leadership position in you know in a in a different domain. uh because I uh you know I I have not always been in the data space.

[00:45:10]

I I've moved from uh you know from credit cards to uh you know core banking products to uh you know apply and buy which I implemented in in another organization and then uh into the data space. So uh I think the advice is would be similar which is a have an open mind when you're going into uh a role like this. go in with the mindset that you're there to learn as much you're there to deliver outcomes. Don't think you're the smartest man in the room. Um you you you know chances are you'll have teams that are smarter uh in different parts. Uh and your job as a leader is to harness that energy and bring the teams together. So never lose your curiosity, never lose your respect and support for for the teams. um because your job uh is to have is to build an integrated you know motivated team which will deliver 100% more for a leader that they respect than uh than they would otherwise be able to do. So I see a lot of leaders uh going into new roles with the frame of mind that you know I got promoted. I'm I'm a really smart person. Um and I'm expected you know sometimes they even put pressure on themselves that I I need to have all the answers and the team wants the answers uh from me. But uh but that but that's you know nothing can be further from the truth because you are you know you're by definition you're new to the role. Uh you've just come in there so you can't expect to have all the answers. There are people there in your teams who are technically more qualified. They'll have better answers.

[00:46:37]

they'll have more experience in in that particular space uh than you have. So uh see your job as providing leadership support and direction for your teams rather than you being the expert. So as long as you do that you're curious and then uh you know I I feel my job always as a leader is is to uh is to ensure the team has the right direction they have the right level of support uh and guidance and providing them with the vision. uh then my job is to get out of their way uh you know and then just just to uh to be available to remove roadblocks uh and support them in delivering the outcomes because the teams are the ones will ultimately deliver for you uh you can't do anything by yourself >> so on that note the su thank you thanks for having here and uh sharing your insights this what I believe those are very relevant to uh what's happening across the industry today uh it was um a pleasure hosting you >> thanks so Right. It's been a pleasure.

[00:47:34]

>> If you found this discussion valuable, please follow and subscribe to Enterprise Tech Talk. And thanks for listening. [music]