Advanced screening technology is crucial in today’s dynamic sanctions landscape
The challenge of balancing cost-effective sanctions screening solutions with the ever-changing regulatory landscape is bigger than ever. As world events progress, as does sanctions regulation meaning that financial institutions need to adopt a risk-based compliance culture to manage the impact. This blog seeks to outline some of the advances in technology solutions which can help firms face these challenges, focussing on name screening.
A high-level summary of the Name Screening Process:
1. Data extraction and analysis:
The first stage in the name screening process is data extraction. In this phase screening technology analyses data to identify potential name matches. This could be through comparing customer onboarding data, or counter-party payment data against internal blacklists or regulator held sanctions lists.
Optical Character Recognition (OCR) can play a significant role in making this process quicker through extracting and analysing data and improving the accuracy of screening hits. Whereas previously, reading handwritten documents was a slow human led task, OCR can scan handwritten documents in real time converting them into digital format, allowing them the text to be mined for potential matches.
Transliteration is another area where OCR brings significant benefits. Previously firms were reliant on staff with multiple language skills to interpret documentation. With OCR, Cyrillic, Arabic and other non-Latin alphabets can be near instantly translated with extremely accurate results. OCR can therefore bring significant benefit in reducing the manual effort required to extract the data to be analysed by screening solutions.
Once extracted, Artificial Intelligence (AI) can support in analysing the data comparing it to blacklists and assessing whether an alert needs to be created or whether it can be suppressed automatically.
2. Alert Generation and Case Investigation:
Once data has been analysed and extracted, alerts are then generated for potential sanctions hits. Assessment by the 1st line of defence determines the validity of an alert.
Robotics Process Automation (RPA) can aid this process, saving time by gathering customer information, or undertaking negative news screening. This information can then be pre-populated into a case narrative.
RPA is also capable of conducting a proportion of first level alert reviews which are resource intensive. Through automating sequential steps, RPA can compare name matches against multiple sources to determine whether a full investigation is needed. This technology can reduce the number of incorrect escalations, saving cost and resource. These savings can enable compliance teams to target their efforts on cases which are deemed of higher concern.
3. Escalation and Reporting:
Where cases are referred to the 2nd line of defence, once the investigation has been completed, a decision to escalate the case to the MRLO is required. Whether a case is escalated or not, reporting of case outcomes and collecting management information (MI) is fundamental for robust compliance frameworks.
Natural Language Processing (NLP) can provide benefits in automating aspects of MI reporting. NLP can help financial institutions harness the data they have by text mining investigation reports categorising references to key words, such as countries, creating additional data. This new data allows financial institutions to have access to richer MI, which can better inform risk assessments allowing them to reallocate resources as appropriate.
Data from false positives can be utilised by Machine Learning to automatically suppress future alerts before they are even generated where degrees of similarities are identified. The extent of which can be suited to the firm’s risk appetite. The benefit of such allows firms to re allocate resources where needed.
4. Outcome Management and Record Keeping:
Once alerts are managed and the outcome data is stored, Unsupervised Machine Learning (UML) can analyse this data to identify hidden or unusual patterns which could emerge into a future risk. UML provides insight that humans alone could not identify. This allows financial institutions to identify emerging risks quicker than with traditional horizon scanning, enhancing their existing risk management frameworks.
The examples of technology mentioned above provide a high-level summary of some of the potential benefits financial institutions can harness from enhancing their sanctions screening technology.
BCS, Part of Accenture has an established Financial Crime Team, with a network of skilled professionals. We can work with your organisation to advise you how best to take advantage of these solutions by identifying areas of inefficiencies to improve your current sanction screening processes.
For more information about our Financial Crime consulting services, please contact: Emma Jordan, [email protected]
Sources:
Operational Transformation of Anti-money Laundering Through Robotic Process Automation | Accenture
client-screening-next-gen-approach.pdf (ripjar.com)
https://www.rawcompliance.com/uploads/6/5/5/4/65548829/pelican_sanctions_datasheet.pdf
https://www.slideshare.net/Enigma_Technologies_Inc/machine-learning-for-sanctions-screening
https://aimagazine.com/ai-strategy/how-ai-can-help-navigate-the-new-landscape-of-sanctions
