Partnering with Absa Group, SymphonyAI has been shortlisted for Technology Partner of the Year
SymphonyAI financial services and South African bank Absa have been shortlisted for the International Compliance Association’s Technology Partner of the Year award. The shortlisted joint project is for using AI to improve transaction monitoring processes to reduce false positives and discover new potential financial crime
The ICA Compliance Awards celebrate the outstanding achievements, collaboration, and innovation of the compliance and financial crime prevention teams across the globe. The official awards ceremony will be held on 27th June 2024.
The submission for Technology Partner of the Year entry
The joint Absa-SymphonyAI submission for the ICA Compliance Awards focused on using AI to improve productivity and risk reduction.
The two companies highlighted key goals for the project:
- Improve alert quality – reduction of false positive alerts, which would be measured against current alerts, while continuing to identify all known suspicious activity reports (SARs).
- Improve new risk detection – Ability to detect new risks faster using the AI that previously would have taken longer to identify and implement by Absa’s current transaction monitoring rules
- Segmentation/transaction monitoring system (TMS) tuning – options for tuning transaction monitoring rules based on Sensa Customer Segmentation.
The teams used masked data from one country covering two years. The Absa-SymphonyAI project then tested the software’s capabilities further by including data from another country.
The project ran smoothly and to schedule with both sides working well together. Absa offered fast, efficient feedback throughout the testing phase, and impressively provided new data quickly, further enhancing findings. The result was a harmonious experience with a lot of optimism for further research and collaboration in future. Absa has already agreed to rollout the software to include three more countries.
The outcome of the proof-of-concept project was extremely positive with results far beyond Absa’s expectations. The results persuaded Absa to contract with SymphonyAI for further collaboration to fully implement the solution and roll out to other jurisdictions.
The project results: positive alerts reduced 77% and 10.5% new risk identification hit rate
SymphonyAI reduced false positive alerts by 77% while also capturing all suspicious activity reports (SARs) as found by Absa using its current transaction monitoring system. Cross validation simulation testing expected a reduction of between 71.9%-81.6%, with 99% confidence.
By sampling 6 months of customer data, SymphonyAI generated new risk models to improve risk detection. Many potential new risks were discovered, which were filtered down to the 200 highest scoring risks based on transaction amount, average score, and maximum score. Absa confirmed 21 new risks were identified quickly versus a slower manual adjustment of rules or manual reporting methods. Alongside this, the new risk identification hit rate of 10.5% was significantly more effective than using rules alone.
During results analysis, roughly 40% were resolved as false positives and learnings from this can be applied to ensure future models are optimal. Additional business information can also help models to not alert on non-suspicious transaction movements. With unmasked data, SymphonyAI expects each set of results to improve even further.
Absa currently uses 15 segments based on account type. SymphonyAI’s initial analysis, based purely on transaction behavior (as data was masked), resulted in four main behavioral segments with 40 sub-segments – a much more granular view of customers and risk.
The future of the partnership
Looking forward, the outcome of this project has seen three more AI proof of concepts planned in areas such as entity resolution and watchlist management, showing commitment to continue the AI journey from both sides. This includes the use of AI to track change in behavior, the benefits of which will see the AI continuously improving.
Absa has also taken part in our customer insights workshop and shared feedback with other EMEA clients, using the team’s own knowledge and experience to help support a more robust and effective financial crime prevention ecosystem.
Because of the success of this collaboration, SymphonyAI hopes for wider adoption of standards for the industry from which everyone will benefit.
Nic Swingler, Head of Financial Crime at Absa said: “Absa has always prided itself on being a leader in the finance industry. We take risk management extremely seriously and by working with SymphonyAI, we are improving our productivity, becoming more effective and reducing risk in an innovative and sensible manner. This will benefit the bank for many years to come.”
Judging panel for the ICA Compliance Award
The ICA Compliance Award judging panel is made up of a diverse range of highly influential and respected figures within the compliance and financial crime prevention community.
These include Neil Whiley, Director of Sanctions at UK Finance; Lisa Bennett, Legal Compliance Director at Mastercard; Vita Danta, Deputy CEO, Board Member and Chief Compliance Officer Europe at @Revolut; and Lizy Purdy, Senior Manager, CIB and LBCM Markets Shared Services CCOR, Markets, Risk at Lloyds Banking Group.
ICA’s President, Pekka Dare commented:
“On behalf of ICA, I would like to personally thank everyone who entered the ICA Compliance Awards. It gives me great pleasure to congratulate SymphonyAI and ABSA Group on being shortlisted as a finalist in the Technology Partner of the Year category. All of the entrants represent the very best of what our industry has to offer, and I am looking forward to welcoming all our finalists to the awards ceremony on the 27th June.”
The selected winners will be announced during the awards ceremony at the Park Plaza Westminster, London.
Learn more about SymphonyAI financial services.
