BANKING SECTOR
Compiled by Yamini Sequeira
BANKING BEYOND THE BRANCH
Dinesh Thomas takes note of the role played by AI in transforming banking
Q: How do you see the role of digital banking evolving?
A: I see banking evolving into a natural part of everyday life rather than something you consciously do.
Everyone will have their own AI assistant, powered by some level of artificial general intelligence, which connects to specialised banking and financial models. These in turn connect to financial service providers to fulfil customers’ needs. Banking will become ambient and invisible.
Q: How has macroeconomic volatility influenced the pace and priorities of digital transformation across the sector?
A: It’s been somewhat of a paradox. The volatility forced people towards digital banking out of sheer necessity but we remain a very cash driven economy – and a rebounding economy doesn’t always lend itself to changing that.
In addition, we have lost a fair share of tech talent to migration, which has slowed the pace at which banks can build and deploy new capabilities.
Q: Given the rise of digital wallets and fintech solutions, how can banks remain relevant while adhering to frameworks?
A: I don’t think banks need to out innovate fintechs; that’s the wrong race to run. What banks have that fintechs can’t replicate is trust and regulatory standing.
Rather than competing feature for feature, the smarter play is for banks to see themselves as the regulated rails that every wallet eventually has to plug into. When things become uncertain, people gravitate towards whoever is licensed and accountable.
So the real opportunity lies in collaboration, not competition.
Q: What do you see as the main challenge for banks in transitioning to more agile cloud based infrastructure?
A: For most banks, the optimal approach is a well architected hybrid model. Cloud migration isn’t a single event; it’s an ongoing journey.
To get that right, managing resources and latency takes real engineering discipline, which is made difficult by how scarce that expertise has become locally.
The regulatory dimension of the Central Bank of Sri Lanka’s requirements around data sovereignty and customer privacy should shape cloud strategy from the outset. Cloud costs being largely dollar denominated also adds pressure, given our currency volatility.
Q: How should banks strike the right balance between innovation and risk mitigation in a digital first environment?
A: This is an important debate but the conversation has shifted. Banking controls were built around the assumption that forgery and impersonation were difficult but that’s no longer the case.
Video KYC can be defeated with deepfakes, call centre verification with voice cloning and credit approvals with AI generated documents. Even signature based authentication faces the same reality.
Maintaining legacy verification processes isn’t the conservative choice it once was; standing still is now the greater risk.
Q: How will AI and machine learning reshape customer engagement and credit risk assessments?
A: Conversational artificial intelligence is what excites me most on the engagement side. Messaging is so deeply embedded in how Sri Lankans communicate that it could genuinely leapfrog traditional app based banking for many people.
Fraud detection is where AI is already quietly delivering results. On credit risk, augmentation is the honest story right now with AI handling the straightforward decisions and reducing workloads, while humans stay in the loop where judgement and ‘explainability’ matter.
But none of this works without first tackling the underlying data quality issue.
Q: What infrastructure investments do you believe are essential for banks in Sri Lanka?
A: Infrastructure investments must be tied to business goals, which is the starting point.
In general terms however, the themes I’d focus on are open application programming interface (API) and data infrastructure, cybersecurity, and anti-scam and fraud management.
Each bank must assess where it sits on the maturity curve and what its business strategy requires next. No two banks are at the same point.
Q: How can banks leverage data and analytics more effectively to personalise services, while maintaining ethical standards and customer trust?
A: Personalisation in banking has to be handled very differently from sectors such as retail. When done poorly, it feels less like a helpful nudge and more like surveillance.
Our regulator offers a useful test here, stating that client data should only be used in ways customers would reasonably expect. Simple things done well can go a long way: a UX that quietly adapts to how you use the app or a timely prompt for a bill you pay every month. That feels like the bank is working for you, not watching you.
We have lost a fair share of tech talent to migration, which has slowed the pace at which banks can build and deploy new capabilities
Transparency around data use then becomes the final trust builder.
Q: How should banks expand access to digital financial services among underbanked and rural populations?
A: For digital inclusion to really work, it must be built around solving genuine problems for underserved communities and structured as a model that makes business sense at the relationship level, not merely the transaction.
Banks must acknowledge that they can’t do this alone: the final mile requires a genuine ecosystem.
Government, community institutions, fintechs and rural networks all have a role to play, because the trust and last mile reach that inclusion needs already exists in those communities, not in bank branches.
