EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
In this episode of the MENA Financial Services Insights podcast, host Mayur Pau is joined by Partner at EY Omair Afzal to explore how data and artificial intelligence (AI) are shaping the future of financial services in the region.
The discussion begins by distinguishing the roles of data and AI, highlighting why organizations must treat them as distinct but connected capabilities. Omair outlines the foundations of data management, data technology and data innovation, and explains how these elements enable advanced analytics and AI use cases.
The conversation then turns to the practical application of AI across financial services, from predictive modeling and fraud detection to credit decisioning and customer relationship management. Omair shares perspectives on where value is already being realized and where organizations continue to face challenges, particularly in applying generative AI effectively.
Listeners will gain insights into the common misconceptions around automation, return on investment and the impact of AI on jobs. Rather than focusing solely on technology, the discussion emphasizes the importance of redesigning processes and investing in people, and skills to fully realize the benefits of AI.
This episode offers a pragmatic view on how financial services organizations can move beyond experimentation and build sustainable value from data and AI.
Key takeaways:
Understand how data and AI differ and why both need distinct strategies
Learn about the core components of effective data foundations in financial services
Discover where AI is already delivering measurable value across banking operations
Find out why process redesign matters more than isolated AI use cases
Gain insights into how workforce skills and culture influence long-term AI success
Mayur Pau:
Welcome to the EY MENA Financial Services Podcast, where we explore how banking is evolving in the Middle East.
I am your host Mayur Pau, and I lead the EY MENA Financial Services business.
Today, we will be covering a hot topic around data and artificial intelligence (AI) in the financial services sector. This topic stretches well beyond a “buzz’” word in the financial services sector with a significant number of use cases developed and implemented across different areas such as fraud and customer relationship management across MENA based banks and insurance companies.
I would like to welcome Omair Afzal, one of data and AI partners at EY focusing on the finance and treasury operations space.
It is great to have you with us, Omair.
Omair Afzal:
Thanks Mayur, happy to be here.
Pau:
We have to start with the title, Data and AI. Should these be placed in the same title or are they now significantly diverged schools of excellence.
Afzal:
When I started in a leading UK bank data team in 2007 it was very much a converged and linked science. The work spanned from basic data preparation in a spreadsheet, to basic BI, to advanced data analysis to building models to hedge complex risk positions. However, the same team does not carry out the same breadth of activities anymore. And while it is important to understand the critical link between these two areas — they need to be approached separately with a clear distinction of the objectives they are serving.
Pau:
That has really opened a host of important points around the evolution of data and AI since the start of your journey in this area. So let us start by clarifying the focus of each item so that we can make the necessary distinctions. What does the data component entail?
Afzal:
Data focuses on providing the right foundations for a number of areas across financial services organizations — this will cut across data management, data technology and data innovation.
The last of those is an area that transcends and provides the segue into AI.
However, let us focus on the data side of the boundary. Data management is an area that many organizations and in particular banks have been focusing on for many years — regulators have also had a huge focus on this if you look at directives such as the BCBS239 and regional guidance such as the National Data Management Office (NDMO) in KSA and the data management framework in the UAE.
These focus on the provision of good controls, good quality data and embedding the right governance processes.
On data technology, which historically was known as enterprise data management (EDM), involves having the right technology and also having the right skills to manage it. The selection of tooling is an area of contention that overlaps with the Chief Information Officer (CIO), e.g. whose responsibility is it to select the actual tools. But management is a key area around being able to extract the right data, transform data sets that are fit for the end user, whether it is in reporting approach or for developing analysis that can range from simple BI to more advanced techniques.
Finally, data innovation or value is what I feel sits on the data side of the Venn Diagram (if we imagine it as that) and is largely BI and financial planning and analysis (FP&A), possibly a little simplified – but this is essentially servicing the organization’s needs – from basic dashboards to more advanced models to support areas such as credit decision-making, operational oversight, and other similar functions.
Pau:
Interesting, you mentioned a Venn Diagram which includes data management, data technology and data innovation. So, what sits on the AI side and what are the overlaps?
Afzal:
It is probably easier to start with what AI is, this is an emerging and fast-developing area, and it is important to stay agile with the agenda in terms of mindset and action.
But as many in our audience know, AI has been around for many years. Even in my early years, we were using RA (i.e., regression analysis) to do predictive modeling around certain market conditions and the impact it would have on derivative positions.
So, it ranges from more basic machine learning, deep learning and modeling that looks to develop predictive models — fraud detection may be an example in the financial services sector.
Currently, natural language processing (NLP) or more broadly, generative AI (GenAI) — which has really captured the audiences with the likes of LLPs, GenAI tools and others — is the next step of the journey.
Now, we also have the buzzwords around Agentic AI and AI agents.
What is clearly on the side of AI is use case definition (or not, and we will touch on this approach later), use case development, use case deployment, lifecycle management and benefits monitoring outcomes and KPIs.
What is overlapping between data and AI is advanced analytics (which could employ machine learning (ML) and deep learning (DL) methods including some elements for data preparation).
AI is not RPA, which is far from AI, building Application Programming Interface (API) integrations or other methods of STP automation, actually. And it is worth highlighting, that on the broader automation path – it is a combination of these methods including AI that deliver more value.
Pau:
That is a great foundation for the discussion and it is interesting to learn that RPA is not a form of AI. If we pivot to the market, where is the value being driven from AI?
Afzal:
There is definitely value, but there is going to be more to come in the future.
Where is the value — I think an area that is very mature is around propensity modeling through machine learning and deep learning methods. Many banks have demonstrated value and one that we have worked closely with in the region is delivering over US$100m in robust enterprise value and another digitally oriented entity in Singapore has publicly stated over US$750m in revenue generation owing to predictive AI models.
From a control perspective, pattern recognition around fraud or contributing to credit decisioning are emerging areas that are fast improving efficiencies and delivering value not only from a cost perspective but also a revenue-loss perspective.
Natural Language processing NLP is still on the exploratory stage. I believe customers still hate being spoken to by agents (and I speak as a customer myself here). However, it is an area that is fast improving and I think it will have a sweeping impact in the future. I currently believe the application of it is the area that is currently lacking.
Pau:
Okay, how can it be applied better from a Customer Relationship Management (CRM) perspective?
Afzal:
I feel like the fundamental approach of use case identification and development is potentially flawed and needs to evolve.
I mean this in a very simple way — what I see many financial services organizations doing is identifying use cases in a “patch” application sort of way on existing processes and then trying to demonstrate Return on Investment (ROI) on them — which I have to say, does not work.
When we had the advent of mobile applications — we did not say “how can we utilize a mobile app for multiple parts of the process?”
We fundamentally redesigned banking or our services around the technology — of course it can always be better, but at least as a principle, it works.
So why does AI have to be different to that. I do not believe it should — we should identify different types of automation or value that can be generated from AI — chatbots, document generation, predictive modeling and redesign our processes around technology.
And the ROI will look after itself.
Pau:
I like the point of ROI taking care of itself, although it will take some convincing of CIOs and other stakeholders on this point. We must cover all the noise in the market with respect to how AI will impact jobs. What is your view for the financial services sector?
Afzal:
Now you are going to get me into trouble.
First, I will start with a more sobering view. In the past, we have had all sorts of predictions around new technologies — around 2013 Robotic Process Automation (RPA) was going to have the huge impact, since then we have had blockchain and how it was going to do away with ledgers and accountants and the list could go on.
I do not need to say how some of those predictions have not come to fruition.
However, I do believe AI is different, and it will have a significantly more revolutionary impact.
It already is in some ways and the key word is — productivity.
We had the same revolutions in other sectors — industrial revolution, farming, etc., and those were largely around how we automate manual physical processes. I believe AI will revolutionize an area that transcends more into the “service” sector and the thinking tasks surrounding the space.
I will conclude on this question — without giving crazy predictions — but by tapping into the scientist in myself and simply by saying that humans have been around for billions of years, and the fundamentals of life have not changed, that much. So, I think AI will massively improve productivity — which will allow us, as a generation, to do even more great things, but it will not change the fundamentals of life. Hope and positivity.
Pau:
Let us then conclude with the “How”. How do we see financial services organizations increase the benefit of AI and what will be the differentiating factors.
Afzal:
This is the one area that I think about the most as part of my role.
And it is so easy to sit here and give a host of factors and reasons that will essentially dilute the answer. But if I was to pick asingle differentiating factor it would be people.
To redesign and retrain what we know as the workforce to become data-centric, to become AI-centric, and that is what will force the organization to form around this great concept.
We are seeing it already, both in our region and around the world, leading banks significantly increasing their AI skills, training programs, etc.
And it is those foundations pillared on “people” that I think will become the differentiating factor, if it is not already the case.
Pau:
Thanks for sharing these great perspectives, Omair. We covered a range of topics today from Data and AI covering the data components around data management, data technology and data innovation. The session then moved into AI developments, use cases and practical ways in which it is already applied in the financial services sector.
Thank you, and I appreciate your time today.
Afzal:
Thanks, Mayur.
Pau:
That’s it for today’s episode on data and AI. Stay tuned for our next episode in the MENA Financial Services podcast series available on ey.com, where we will dive deeper into other financial services topics.
Thanks for listening.
