“Up to Rs 1 million loans secured within 48 hours without any human intervention”

AI in the Financial Sector

• “The pace of technological change requires attracting global expertise, as Singapore and Dubai have done”
• “We developed a Creole-language emotional intelligence model ”

During the Gen AI Summit for Sustainable Development 2025, senior leaders from Cim Finance, Absa Bank, and the Mauritius Commercial Bank engaged in a far-reaching panel discussion on the rapid transformation of the financial sector through artificial intelligence. Moderated by Daniel Essoo, CEO of the Mauritius Bankers Association, the conversation explored how Mauritian banks are integrating AI into customer journeys, credit assessment, operational processes, and cybersecurity. It also examined the growing importance of data and technology in climate finance. The discussion revealed a financial ecosystem that is already undergoing profound change, with digital onboarding, automated lending, emotional intelligence models in Creole, and advanced payment solutions emerging as tangible innovations shaping the future of financial services.

Daniel Essoo – Our panel today brings together senior figures from the banking and financial services industry. They will help us understand developments that are often not visible from the outside—how banks are being transformed internally, how technologies are reshaping finance, and how sustainability and climate considerations are increasingly embedded in financial processes. We will begin by exploring each institution’s innovation journey and then move into climate finance and the role of technology in building a more sustainable financial ecosystem. Let me start with our colleagues from CIM. Sudheer, your institution has a strong track record in innovation—you were among the first to launch a green bond and advanced automation solutions. Could you share CIM Finance’s journey in embedding technology and AI within operations?

Sudheer Prabhu (Group CTO, CIM Finance) – At CIM Finance we approach innovation through three core lenses: customer experience, operational excellence, and cybersecurity. Our philosophy has always been to use technology to solve real problems rather than adopting technology for its own sake. A key example is our recently launched product MoPay, a 0% interest, five-instalment digital credit solution that is fully AI-driven and entirely paperless. The process begins when a customer enters a shop, scans a QR code, and receives an offer. The system captures a photo, identifies and matches the NIC, and automatically processes risk scoring. Customers can walk out with approved credit in just 15 minutes.

Another important innovation concerns our collections function. Since collections involve emotional nuances and communication patterns, we developed a Creole-language emotional intelligence model. It analyses audio clips, classifies emotions such as calmness or anger, identifies linguistic cues, and helps our agents adapt their communication strategies. This has proved especially valuable given the diversity of Creole speech patterns.

In operational excellence, we now use GenAI to analyse financial documents. Previously, analysts manually extracted balance-sheet figures, inserted them into spreadsheets, and calculated ratios before writing credit memos. Today, analysts can upload the balance sheet into the model, which instantly generates the required ratios and the basis for the credit assessment. This has significantly increased efficiency.

Finally, cybersecurity is a critical part of our AI journey. Machine learning models constantly monitor transaction patterns to detect unusual behaviour, potential intrusions, and fraudulent documents. Alerts are raised in real time for supervisory review. This proactive detection helps us act before any threat escalates. These are some of our most impactful AI-driven developments.

Daniel Essoo – Thank you, Sudheer. Several elements stood out. First, the end-to-end automation in customer journeys – from identity verification to risk scoring—represents a major shift. Second, the emotional intelligence model in Creole is particularly impressive since it addresses local language realities. And third, the cybersecurity applications show how deeply AI is now integrated into risk management. Let me now turn to Absa and invite Sandeep to share their experience. You were early adopters of AI with the introduction of Abby, and you’ve continued accelerating digital transformation. What has your journey looked like?

Sandeep Mohapatra (Digital Transformation Leader, Absa Bank) – Thank you, Daniel. At Absa we follow a principle: your story matters – the story of our colleagues, our customers, and our communities. We therefore start with a customer problem and allow technology to provide the solution. This means we remain technology-agnostic and use AI only where it is the most effective tool.

We asked basic but important questions. Why must customers physically visit a branch to open an account? Why should they wait in queues, sign paperwork, or submit documents only during office hours? Why can’t someone open a bank account digitally on a Saturday afternoon? These questions led us to develop Abby, and eventually to create Mauritius’s only 100% digital onboarding process using liveness detection, document intelligence, and now integration with the government’s CKYC system.

The next major challenge was digital credit. We believed it would be straightforward, but we discovered how complex income structures, salary slips, bank statements, and MCIB reports could be. Building the intelligence to validate income, detect fraudulent documents, reconcile bank statements, and analyse over 200 credit ratios required deep learning models and countless hours of calibration. Today, customers can submit documents online and receive a provisional credit decision within minutes.

“Over the past four months alone, more than 1,000 applications were approved through the engine”

Our latest development, Spark Business, addresses SME payment challenges. SMEs often rely on cash because POS machines are costly, reconciliation is complicated, and small merchants lack ERP systems. Spark Business converts a smartphone into an omnichannel payment device capable of accepting QR payments, tap-to-phone card transactions, issuing digital receipts, sending remote payment links, and providing simplified reconciliation. This gives small merchants the same technological power previously reserved for larger retailers, at a fraction of the cost.

Daniel Essoo – Thank you, Sandeep. Three things resonated strongly. First, the problem-driven approach; technology follows customer needs, not the other way around. Second, fully digital account opening is transformational, especially when many people still complain about long account-opening delays. And third, the merchant-facing innovations you described directly support the transition to a more digital, less cash-dependent economy. Let me now turn to Harry from MCB, who has been working on an exciting project related to SME lending. Harry, could you walk us through what you’ve been developing?

Harry Coolen (SME Banking Manager, MCB) – Thank you, Daniel. Mauritius may be small, but it has always been innovative. Years ago, MCB introduced the Mr. BEST card, which helped propel the shift from a cash economy to digital payments. Today, MCB Juice and JuicePro are part of everyday life and have even influenced local language.

Building on this digital culture, we embarked in 2022 on developing an internal credit-scoring engine for SMEs, embedded within JuicePro. This engine uses machine learning and combines internal and external data. After three years of development and refinement, we have reached a significant level of maturity. Over the past four months alone, more than 1,000 applications were approved through the engine. Volumes this year have already surpassed those of the previous year, and 30% of SME credit applications now go through this automated channel.

Most importantly, SMEs can secure Express Loans and Express Overdrafts of up to Rs 1 million within 48 hours, without any human intervention. The process is fully automated, including e-signatures for contracts. In the last four months, we have already disbursed Rs 1.5 billion through this channel. We are now preparing to expand exposure limits and integrate additional products such as trade finance. The momentum is strong, and the feedback from SMEs has been extremely positive.

Daniel Essoo – Let me highlight the magnitude of this achievement. For years, all banks combined struggled to reach Rs 1 to 1.5 billion in SME lending annually, even with government support through guarantee schemes. What you’re describing is a quiet revolution: MCB alone has reached Rs 1.5 billion year-to-date, with roughly Rs 500 million processed without any human intervention. This is a major leap in both efficiency and customer experience.

Before we move to the final question, I want to briefly address the climate dimension. As all panellists mentioned, AI relies heavily on data. Climate finance is no exception. Traditionally, climate-related datasets in Mauritius were not easily accessible. However, Government decided in the last budget to release significant climate-risk information previously held within various ministries. Flood-risk maps and land-drainage zones are now publicly available, meaning banks can integrate these into credit assessments. Work is also underway on a national climate-risk map, supported by the Agence Française de Développement. AFD has already completed extensive studies on flood exposure, wind patterns, and coastal erosion – crucial for banks with exposure to hotels and coastal development. The ultimate goal is to create a unified, accessible climate data ecosystem that AI tools can incorporate into risk-modelling processes.

With this context in mind, let me move to our final round. If an AI genie granted each of you three wishes to accelerate the national AI journey, what would they be?

Sudheer Prabhu – My first wish would be to make MCIB data accessible through APIs so that every bank can integrate it into automated credit processes. Second, I would like to see a single, nationwide credit model maintained centrally and accessible to all financial institutions. This would reduce duplicated computing resources and environmental impact. Third, a centralised address-verification system, also accessible through APIs, would standardise and streamline verification across all institutions.

Daniel Essoo – For context, MCIB is our national credit bureau. In many countries, banks connect directly through APIs, receiving structured data and analytics. In Mauritius, MCIB still issues a PDF, making automation difficult. So, Sudheer’s wish reflects a long-standing sector need.

Sandeep Mohapatra – My first wish would be the availability of deep technical talent. We have bright students, but the pace of technological change requires attracting global expertise, as Singapore and Dubai have done, while continuing to build local talent.

My second wish relates to a level playing field in regulations and infrastructure. For example, CKYC is only partially implemented – mobile-number verification should be integrated as a second factor. We also need account-to-account payment rails and a domestic card switch to reduce dependency on expensive international networks.

Finally, I wish for a comprehensive digitisation of government services. Many public processes still require customers to queue, pay in cash, or use cheques. For AI to reach its full potential, public services must be modernised in partnership with the private sector.

Harry Coolen – My wishes focus on climate and data. First, I would like a unified national climate-risk data hub so banks can assess exposure across various sectors. Second, access through API to energy-consumption data, which we currently process manually. This would help us identify customers who may benefit from shifting to renewable energy and support our green-finance initiatives. Third, I would like climate and energy data to feed directly into AI-driven credit models, allowing banks to strengthen both financial and environmental decision-making.