QuantaVerse’s AI-powered AML solutions secure highest possible ratings from independent model validation firm
- Artificial Intelligence , Security and Compliance , AML and KYC
- 23.07.2020 11:25 am
QuantaVerse, which uses AI and machine learning to automate financial crime investigation processes, engaged an independent model validation firm to ensure the regulatory rigor of its AI and machine learning platform. Leveraging its proprietary modeling framework to thoroughly analyze QuantaVerse’s AI solutions, the firm delivered a comprehensive model validation.
Each solution, including the QuantaVerse Alert Investigator and QuantaVerse False Negative Identification and Investigator, were given the highest possible marks, proving that its AI models are effective and work as intended. QuantaVerse, having been in production in client environments for several years, can now lay claim to being the industry’s only independently validated comprehensive technology platform to deliver AI for AML.
“Models are used throughout the banking and financial services space, particularly for determining risk. As such, bank regulators issue guidance advising periodic model validation,” explained David McLaughlin, CEO and Founder of QuantaVerse. “We encourage banks and financial services companies to review our validation results as part of their regulatory and due diligence requirements. Our validation results are proof positive that the QuantaVerse platform delivers on the promise of AI and machine learning for AML/BSA compliance efforts.”
When presented with input data such as negative/adverse media or transaction data, QuantaVerse’s models produce a value, or observable, which is used to determine the presence or prevalence of financial crime. The models vary widely depending on the nature of the observable they are trying to infer, from natural language processing techniques to neural networks to specialized graphing techniques.
Critical to characterizing the risk associated with an entity or transacting party, the validation focused on model-driven observables relating to profile, monitoring, reputation, and intent. The model validation firm developed generic data sets to exhibit specific characteristics and behaviors that were then ingested and processed through the QuantaVerse platform and, specifically, its AML and screening solutions. The output from the AI solutions was then sampled and reconciled with the input data. All QuantaVerse governance and processes were examined with the highest possible marks.
The validation firm utilized the following methodology to complete its analysis:
Developed detailed transaction and entity test data as input to exercise the model
Processed multiple iterations of simulated transaction and entity data
Analyzed outputs of the system to validate the results against expected values
Logged findings and observations on each observable
A copy of the QuantaVerse model validation report summary can be requested by visiting: www.QuantaVerse.net/contact.