Before AI Comes Clean Data

By Mahesh Paolini-Subramanya, Founder and CTO, BKN301

Artificial Intelligence (AI) is becoming integral to financial services, involved in scoring transactions, customer interactions, and risk management. However, the core operating system for these applications is data, not AI itself. Many financial institutions are using outdated systems that do not support real-time intelligence.

AI's Rapid Expansion and Data Constraints

According to PwC research , over a third of CEOs in the Middle East have implemented AI in their products and services. Despite this, AI performance is plateauing due to data constraints rather than model limitations. Although 70% of CEOs report having AI roadmaps, results are inconsistent due to fragmented data.

Similar to unreliable APIs affecting software development, unstable data environments hinder AI's potential. Models require consistent data schemas and reliable lineage for effective output. Many institutions still operate on systems designed for transaction processing, which are inadequate for real-time AI applications.

Real-time AI systems are under regulatory scrutiny as they influence critical decisions like credit approvals and fraud detection. Authorities in the UAE emphasize strong data governance and model risk controls. Additionally, customer expectations for fast and accurate digital services are increasing, adding pressure on financial institutions to improve data transparency.

For the Middle East to achieve global AI leadership, financial institutions must modernize their data infrastructures. Reliable data is essential for automation and scalability. Addressing weaknesses in data structure and governance is crucial for moving AI from pilot projects to everyday operations, ensuring sustainable AI leadership for the region.

Based on reporting by techround.co.uk.