AI In Finance Round II—Corporate Winners

Arab Bank

Arab Bank has revolutionized its Corporate and Institutional Banking lead generation process with the deployment of an automated business-to-business solution tailored for trade finance. This state-of-the-art AI platform leverages natural language processing (NLP), GenAI, and LLMs to analyze transaction data and identify high-potential leads with exceptional accuracy.

Key features of the solution include dynamic data cleansing and recognition capabilities, enabling the accurate identification of misspelled company names and the unification of matching entities. Through its automated dashboard, the platform provides a clear visualization of analyzed data, highlighting both untapped trade opportunities among current clients and new prospects to be pursued. The automation of this process has eliminated manual lead identification, significantly increasing operational efficiency and deepening insights into client relationships, while also enhancing the precision of business planning.

The integration of LLM and GenAI technologies further refines data interpretation, manages language nuances, and predicts customer behavior, enabling precise lead prioritization. Since its implementation across all PLC countries, the solution has accelerated the identification of high-potential leads and contributed to a notable increase in Arab Bank’s share of the trade finance cycle. This initiative underscores Arab Bank’s commitment to innovation and supports the bank’s strategic growth objectives.

Banco do Brasil (via Capgemini)

Banco do Brasil integrates AI and GenAI into its corporate payment systems, primarily for applications such as fraud monitoring and AML analysis and to provide financial intelligence to its small and micro business clients. The institution currently employs over 700 distinct AI models across its array of operational solutions.

An AI-powered conversational agent, called Aladin, assists AML analysts by synthesizing information from regulatory documents and public data sources, accelerating case evaluations and ensuring adherence to compliance mandates. The bank is also leveraging its cloud-based infrastructure to enable real-time fraud monitoring, including of transactions processed through Pix, Brazil’s instant payment system.

An AI model for corporate default alerts represents a notable advancement in prevention, mitigation, and control of potential defaults within high-revenue corporate credit portfolios.

By employing sophisticated machine learning techniques and integrating diverse data streams through this model, Banco do Brasil is able to offer an innovative and adaptive approach to risk management, proactively inhibiting nonpayment and thereby safeguarding the bank’s financial sustainability.

Arab Bank

Arab Bank has implemented a robust Risk-Adjusted Return on Capital (RAROC) solution, a critical financial metric calculated at both the group and individual customer levels. This advanced solution enables comprehensive assessments of financial health and risk for corporate clients by leveraging extensive financial datasets.

The RAROC platform segments customers by country, income, and risk profile, delivering strategic insights that support risk reduction and optimized capital allocation. Enhanced data visualization capabilities and intuitive drill-down features allow users to explore specific customer clusters, facilitating adoption and empowering stakeholders with actionable, tailored insights.

This initiative underscores Arab Bank’s commitment to leveraging data-driven solutions to strengthen risk management and support informed decision-making across its corporate banking portfolio.

Bank of East Asia

Hong Kong-based Bank of East Asia has applied AI technologies to a range of internal solutions, positioning itself to compete with its fintech rivals.

An AML solution employs machine learning, historical libraries, and feature analysis to scrutinize transaction flows to detect anomalies, minimize false positives, and boost the efficiency of AML investigations. Call Center Intelligent Assistant uses a pre-trained LLM and a knowledge graph (or smart database model) to quickly find and deliver accurate business information, helping agents answer customers’ questions faster and more accurately.

A locally deployed GenAI tool automatically produces credit reports for business clients. The bank reports that the tool achieves a reference rate exceeding 60% for industry, financial, and management information analysis sections. The bank’s intelligent cash management solution integrates AI, big data-built cash demand forecasting, and reasonable interval models into existing tools, optimizing counter cash management for bank managers.

Citi

Preparation of due diligence questionnaires, including those mandated for RFIs and RFPs, is typically a manual, time-intensive process prone to quality inconsistencies. Citi has developed a novel GenAI solution to address the problem, both for its proposal teams and for SMEs, an innovation it expects will enhance client engagement. 

Part of Citi Services’ sales assist tools, the new solution is positioned as the equivalent for unstructured content to the service data warehousing provides for structured data, automating processing and unlocking new insights. Citi aims to integrate the solution with its existing data assets to fuel further GenAI-powered data products.

Smartstream

Air, Smartstream’s AI-powered, cloud-native platform, aims to turn data into a strategic asset by using AI and machine learning to process, cleanse, enrich, and match data from diverse sources and formats, including unstructured data. Features like a native co-pilot and advanced reporting help users maximize data value, boost accuracy, cut costs, mitigate risk, and strengthen security.

Air ingests information from internal systems, third-party applications like portfolio/order management systems, and databases, supporting all data formats and sizes and automatically uploading unstructured data. The latest version, 9.0, establishes Air as a leading data automation and intelligence platform. Built on modern software-as-a-service technology, 9.0 is globally available, scalable, fast, and cost-effective, allowing immediate customer implementation.

Air’s use cases are now enterprise-spanning, from trade to post-trade, helping to identify and correct errors, clean data, and support regulatory reporting, and its agile architecture extends its applicability to other industry verticals. Version 9.0 introduces two AI/machine learning-enabled modules, Air Data and Air Cash. Air Data automates time-intensive tasks like cross-system checks, pre-settlement error detection, and trade record comparison, and improves internal data quality by flagging inconsistencies in static/reference data. Air Cash quickly handles a range of cash reconciliations, from simple to complex cases.