ANZ Bank
6 min
What our client needed
ANZ Bank provides banking and financial products and services to over 8.5 million retail and business customers, and operate across close to 30 markets.
In the fast-paced world of finance, ANZ Bank grappled with the intricacies of manual Foreign Exchange (FX) trade reconciliation.
Faced with a cumbersome and error-prone process, ANZ sought a transformative solution to streamline operations.
What we gave them
Armed with a cutting-edge approach integrating Artificial Intelligence (AI) and Machine Learning (ML), DCX agency Reason set about revolutionising the FX trade reconciliation process by automating workflows and making use of advanced technologies.
Phase Zero - Establishing the Workflow
In a swift four-week span, we validated the robotic process automation workflow through an interactive prototype.
This crucial phase laid the groundwork, defining the approach, target architecture, and project delivery timeline.
Phase One - Automating Workflows with Chatbots
The initial release marked a milestone with a 50% auto-match rate on FX trades.
The integration of AI-driven chatbots and a dedicated TradeButler web application seamlessly became part of daily workflows, flagging unmatched trades for swift resolution.
- 50% auto-match rate achieved
- TradeButler web application and chatbots seamlessly integrated into daily workflows
- Efficient resolution of unmatched trades with links to supporting documentation
Phase Two - Ingesting and Interpreting Multiple Data Sources
Advancing into the second phase, we implemented regression analysis on new data-sets, pushing the automatic trade match rate to an impressive 70%.
The enhanced functionality of chatbots and the TradeButler Web App facilitated inter-departmental communication, significantly accelerating the resolution of non-matched trades.
- 70% auto-matching of FX trades attained
- 75% faster resolution of non-matched trades using advanced Web App and Chatbot features
Phase Three - Driving Match Rates with Machine Learning
The final phase saw the implementation of new machine learning models, specifically prodi.gy, employing entity recognition and predictive recommendations.
These models autonomously determined whether to auto-match a trade, steadily pushing match rates beyond 90%. Manual processing time witnessed a substantial 75% reduction.
What was the impact
Our architectural prowess delivered secure and effective solutions for ANZ Bank, blending adherence to internal policies with operational flexibility.
Docker containerisation streamlined concerns, while an event-driven micro-services architecture facilitated smooth deployments for new features.
- Symphony’s Customer Innovation Award recognised the groundbreaking achievement
- A stellar 90% automated trade match rate through the synergy of AI and machine learning
- Rapid 4-week timeline to deliver a prototype and validate the automation process
- A remarkable 75% reduction in time spent on manual reconciliations
In essence, ANZ Bank's journey into the realm of AI and ML not only overcame manual trade reconciliation challenges, but also set a new standard for efficiency and innovation in the financial sector.