Article by: Marwan Abouzeid, Principal Digital Strategy Consultant, Backbase
If there is one expectation that defines modern banking, it is frictionless service. Customers are no longer comparing banks solely with one another, but with the seamless, one-click conveniences they experience across so many facets of daily life. Research shows that more than 88% of customers prioritize experience when choosing a financial provider, and in the UAE, where digital adoption is high, expectations are even sharper.
UAE banks have made significant progress in response. Account opening, once a branch-bound exercise, can now be completed via an app in minutes, while customer service has shifted from long wait times to real-time, digital-first interactions. Despite these advances, friction has not disappeared. It has simply moved deeper into the process, surfacing in slower decisions, manual handoffs, and fragmented workflows that customers still feel, even if they cannot always see them.
When process efficiency plateaus
This is not a question of effort or investment. UAE banks have invested heavily in automation and digitization, but much of this has focused on optimizing individual processes rather than rethinking end-to-end workflows.
Lending origination is a clear example. A typical SME loan application involves trade license verification, credit bureau checks, Emirates ID validation, employment confirmation, income assessment, credit decisioning, and often, Islamic product classification.
While many of these steps are impressively automated, the process as a whole still spans multiple systems, teams, and decision points. That fragmentation creates reliance on human coordination. Documents are chased, approvals are escalated, and exceptions are resolved manually across functions. The result is cycle times of 10–20 days for SME loans and 5–7 days for retail lending, with additional delays for Sharia compliance. The components may be efficient, but the overall journey remains disjointed.
The agentic AI misconception in banking
According to IBM, 86% of executives believe AI agents will significantly enhance process automation by 2027. It is therefore unsurprising that banks are looking to agentic AI as the next step forward. However, there is a risk that banks repeat the same pattern of chasing marginal efficiency gains.
If AI is applied to optimize isolated tasks rather than orchestrate entire workflows, the gains will remain incremental. Rather than resolving fundamentally fragmented processes, more sophisticated tools risk adding complexity without materially improving outcomes.
The constraint is not what technology can do within a step, but how effectively those steps are connected. Until front, middle, and back-office processes operate as a unified flow, meaningful reductions in cycle time will remain out of reach.
The build-versus-buy trade-off
The urgency to address this is growing. Digital-native lenders and embedded finance players are already making credit decisions in hours rather than days, forcing traditional banks to rethink how they close the gap.
The instinctive response is often to build. On the surface, developing orchestration capabilities in-house offers control and the promise of tailored solutions. However, in the UAE context, this quickly becomes complex and resource-intensive. In a market where speed is becoming a competitive differentiator, multi-year build cycles risk eroding any advantage before it is realized.
Additionally, integrating with local regulatory and data ecosystems such as AECB, Emirates ID, MOHRE, and the FTA is not a one-time effort. These connectors require continuous maintenance and alignment with evolving regulatory expectations, while governance frameworks must keep pace with increasing scrutiny around AI and financial crime.
Additionally, integration and upkeep drain engineering capacity away from innovation — every sprint cycle spent patching connectors, maintaining custom APIs, or resolving data inconsistencies is a sprint cycle not spent building features that grow revenue. The backlog fills with technical debt instead of product value. This introduces a significant opportunity cost, particularly as competitors accelerate time-to-market using pre-built capabilities, shipping new journeys in weeks while internal teams are still in maintenance mode.
This does not make building the wrong choice, but it does change the equation. The full cost, including financial, operational, and competitive considerations, must be understood before committing to it.
The hybrid model of modernization
A more pragmatic approach could be to balance speed with control through a hybrid model, often referred to as progressive modernization by industry experts. By leveraging a platform as a foundational backbone, banks can accelerate initial deployment using proven, out-of-the-box journeys, reducing time-to-market without unnecessary reinvention.
From there, the model evolves as banks gain insight into customer behavior and operational gaps. Journeys can be extended, refined, and customized to support differentiation, allowing institutions to move quickly at the outset while retaining the flexibility to innovate over time.
Building for the UAE market
However, the speed of transformation alone is not enough. The ability to orchestrate effectively depends on platform credibility, and in the UAE, that credibility is defined by how well local nuances are handled.
KYC and customer due diligence must integrate seamlessly with national identity systems and regulatory frameworks, while financial crime operations need to support real-time monitoring, case management, and reporting aligned with local expectations. SME lending workflows must connect multiple data sources without introducing manual bottlenecks, and call center operations must bridge digital and human interactions without disrupting the customer experience.
Islamic finance represents the most critical test. Products such as Murabaha, Ijara, and Wakala are central to the UAE market, yet many platforms do not support them natively. This often results in workarounds that reintroduce complexity, extend timelines, and increase operational risk. A platform that cannot handle these requirements from the outset will struggle to deliver true end-to-end efficiency.
Governing the next layer of complexity
Once banks begin to address orchestration and platform fit, a new layer of complexity naturally emerges: how to govern these increasingly intelligent workflows. Governance is not a parallel concern, but a necessary extension of orchestration at scale.
As AI agents are introduced into workflows, the challenge shifts from enabling automation to controlling it. Processes must be fully traceable, with clear audit trails capturing how decisions are made and executed. Authorization frameworks need to define the boundaries within which agents can operate, while human oversight remains embedded in critical decision points.
In a market shaped by tightening regulatory expectations, particularly around AI and financial crime, these capabilities are essential. Governance is no longer a back-end requirement; it is a core design principle that will increasingly shape platform selection.
Getting the sequence right
For UAE banks, the advice is simple: resist the impulse to lead with technology and focus first on getting the decision sequence right.
Start with an honest build-versus-buy assessment. This cannot just be the version that fits the initial business case, but the one that accounts for the full costs along the journey, including regulatory connector maintenance, governance infrastructure, engineering opportunity cost, and a market that is not standing still. Once that question is settled, platform selection becomes clearer. The right platform is the one that meets UAE-specific requirements without compromise and delivers production-ready regulatory integration, Islamic finance workflows that need no custom rebuild, and governance that holds up under CBUAE examination.
