The next step in financial inclusion may not be human at all

07/11/2025 02:01 AST

Across the Middle East, financial access has expanded faster than ever before. Millions now use mobile apps to simply and quickly pay, transfer, and track money. In Saudi Arabia and the UAE, mobile penetration rates exceed 95 percent, which enables widespread access to digital banking and payment platforms, reducing dependence on physical branches. In Egypt, financial inclusion has more than doubled since 2016 .

This has been driven in large part by mobile penetration, which now exceeds 100 percent . Individuals across income levels are able to open accounts, transact, and engage with financial services on their own terms. Despite this progress, full participation remains out of reach for many.

People may have accounts, but they do not use them consistently to save, borrow, or plan for the future. There are several factors which contribute to this gap. In some markets, physical infrastructure remains limited. In others, low financial literacy, digital skills gaps, and high service costs continue to block meaningful engagement. For certain populations, including women and low-income workers, cultural and structural barriers persist. The result is a system that is open in principle but often not truly accessible in practice.

This is the gap financial institutions are now examining more closely, and where a new kind of AI, known as agentic AI, is beginning to show its greatest potential for transformative impact.

What sets Agentic AI apart
Agentic AI refers to systems that operate independently within defined boundaries to support users. Unlike earlier models built for single tasks, these agents can process a range of inputs, recognise behavioural patterns, and make decisions based on context. They are designed to step in at key moments to reduce friction and make interactions with financial providers easier to understand and manage.

These capabilities are still developing, but early deployments of agentic AI are already underway at leading global institutions. BlackRock, JPMorgan, and BNY Mellon are using these systems to analyse large data sets, streamline operations, and improve reliability in core banking functions . While these tools have so far focused on operations, they mark a shift in how banks are beginning to trust autonomous systems at scale and are laying the groundwork for customer-facing use cases.

In the UAE, this new model of banking is already reaching the customer. Some banks are deploying agentic systems to enable autonomous decision-making in areas such as personal lending and SME financing. While these innovations reinforce the region's strength in digital banking, the greater promise lies in how agentic AI can address long-standing inclusion challenges.

Three problems Agentic AI can help solve
Financial inclusion has focused on access. Agentic AI shifts the focus to engagement - the area most financial systems are yet to solve. Where traditional banking systems wait for a request, agentic AI can act before a customer asks. This shift matters because it helps prevent the hesitation and uncertainty that cause people to disengage. Agentic systems can detect those points and respond before that engagement is lost.

Supporting users who lack financial literacy or digital confidence
A significant portion of the underbanked population in the region faces challenges related to financial understanding and digital navigation. Many are first-time users of formal banking systems or still rely on cash as their primary monetary model. For those users, digital processes often feel intimidating, and the absence of timely help can cause them to exit before completing a task. Agentic AI can offer clear, step-by-step prompts in everyday language, guiding the user forward without requiring them to seek assistance. When delivered through familiar channels like WhatsApp or through voice-enabled interfaces, financial services become more accessible, helping bridge the gap between simply having access and being able to use it with confidence.

Adapting to irregular income and alternative financial signals
Traditional financial systems tend to serve people with fixed monthly salaries and clean credit histories. But many individuals, including small business owners, freelancers, or workers paid in cash, operate on variable income patterns. This makes it difficult for banks to offer standardised credit or savings products. Agentic AI could help bridge this gap by interpreting behavioural data instead of relying on fixed documentation. For example, it might notice that a customer receives several transfers around the same time each month and suggest saving a portion, or flag upcoming shortfalls based on spending trends. It could also use these insights to support alternative credit assessments, helping people qualify for products that would otherwise be out of reach.

Responding to users in their own language and local context
Language can be a quiet barrier in digital banking. While many platforms default to English, several users across the GCC and North Africa are more comfortable in Arabic, Urdu, or other regional languages. Even for fluent English speakers, the language of financial services can feel technical or unfamiliar. When someone is already unsure about a financial decision, having to navigate a service in a second language compounds the difficulty. Agentic systems that are trained to respond in the customer's preferred language, and translate financial jargon into everyday terms, can reduce hesitation. For users on the edges of the formal financial system, this familiarity can be the factor that keeps them engaged.

Looking ahead
For financial institutions, the task now is to focus less on expanding reach and more on utilizing intelligence tools to ensure a smooth and user-friendly process for the services that already exists. If services can adapt to income volatility, guide users who hesitate, and communicate in the right language and format, they will be used more often and by more people. This is when financial inclusion can begin to take hold. The institutions that understand and anticipate the power of AI agents and set the foundations within their structure to enable them, will define what comes next.

The writer is Group Head of Technology, Transformation & Information, Mashreq

Khaleej Times