Saurabh Kunj
SVP Product Management – Banking
The recent AI guidance issued by the Central Bank of the United Arab Emirates (CBUAE) marks a structural shift in how artificial intelligence is viewed in financial services. AI is no longer positioned as an innovation initiative at the edge of the organisation, it is now firmly within the scope of supervisory oversight.
For those of us responsible for building mission-critical BFS platforms, the signal is clear: AI has moved from experimentation to expectation.
At Azentio, our BFS products are the digital backbone that enable institutions to run, grow and transform their business. That foundation, stable, secure, compliant, enterprise-grade platforms, is exactly where AI must now be embedded. Not as an overlay, but as an accountable, governed capability.
AI under supervision
The CBUAE guidance reinforces several core principles:
- Clear governance and board accountability for AI deployments
- Robust model lifecycle management
- Explainability and documentation
- Ongoing monitoring and risk controls
- Alignment with existing compliance and operational frameworks
- Consumer protection safeguards
In practice, this means that AI in areas such as digital onboarding, AML monitoring, credit underwriting, fraud detection and customer engagement must now be auditable and defensible by design.
This aligns closely with how we think about AI across our product portfolio.
Our three pillars of AI in BFS
When we talk about AI across digital banking, lending and compliance, we anchor it around three pillars.
-
AI as a Catalyst
AI must accelerate outcomes. Whether it is smarter underwriting and risk scoring, fraud detection, predictive insights, agent assist, conversational banking or automated classification, the objective is tangible improvement in speed, accuracy and decision quality.
But acceleration without control is unsustainable. The CBUAE’s guidance makes it clear that efficiency gains must sit within defined governance structures. Catalyst impact must be measurable, and manageable.
This is why our roadmap includes clearly defined AI priorities such as reporting AI, document & analytics AI, product recommendation, smart underwriting and risk scoring, evolving our solutions towards an agentic AI model. Each is anchored to business value and operational ROI, not abstract experimentation, ensuring AI is applied in ways that support stronger decisions, better responsiveness, and more intelligent day-to-day operations.
-
AI with context
Financial services operates within a highly regulated, policy-driven environment. AI models cannot be generic; they must understand product structures, Shari’ah requirements, local regulations and risk frameworks. They must also support fair, non-discriminatory outcomes, especially in areas such as credit underwriting, where bias and weak governance can have direct customer and regulatory consequences.
Our product strategy is built around cloud-native, microservices architecture, platform-first extensibility and API-led digital ecosystems. This architecture enables AI to be embedded into workflows rather than bolted on, with stronger human oversight, clearer accountability and better control over how models are used in practice.
In compliance specifically, our FRAML strategy brings Fraud and AML together into a unified analytics-driven solution with shared data, risk scoring and case management. This unified approach provides contextual intelligence rather than siloed alerts, while also improving traceability and supporting stronger oversight across internal teams and third-party providers.
Regulators are increasingly looking at effectiveness, not just rule coverage. Contextual AI improves signal quality while strengthening governance documentation, case traceability and accountable decision-making.
-
AI that lasts
Banks do not deploy systems for quarters; they deploy them for decades.
The CBUAE guidance highlights lifecycle management, documentation and supervisory oversight. That demands AI capability that is sustainable.
Across lending, we are stabilising, productising and scaling with execution discipline, hardening the platform before expanding feature velocity. AI must follow the same principle. Foundation before expansion.
Cloud-native microservices, modular design and predictable release cadence are not just technical decisions. They are governance enablers. They ensure models can be updated, monitored and refined without destabilising the core platform.
From innovation to infrastructure
The key shift introduced by the CBUAE guidance is this: AI is becoming infrastructure.
It must support institutions in running the business reliably, growing through digital journeys and transforming through modern architecture. That means embedding governance at the design stage:
- Clear model documentation
- Role-based access controls
- Case traceability
- Audit trails
- Risk monitoring dashboards
- Board-level visibility
This is not about slowing innovation, it is about scaling it responsibly.
What financial institutions should be doing now
For banks and NBFCs across the UAE, the next steps are practical:
- Inventory existing AI use cases
- Map them against governance expectations
- Identify documentation and oversight gaps
- Strengthen model monitoring processes
- Align board reporting with AI risk exposure
- Embed lifecycle thinking into new AI initiatives
Institutions that move early and structurally will gain advantage. Governance will increasingly become a competitive differentiator, particularly in compliance, digital onboarding and risk analytics.
The opportunity
The UAE continues to position itself at the forefront of financial innovation and clear regulatory guidance provides confidence to scale AI responsibly.
For product teams in BFS, the responsibility is clear. We must design AI that accelerates performance, operates within regulatory context and stands the test of supervisory scrutiny.
Catalyst in impact. Grounded in context. Built to last.
That is the standard AI in banking is moving toward, and it is the standard we are building for.
Saurabh Kunj
SVP Product Management – Banking