Overview
What is Adverse Media Monitoring
Evan (formerly Evelyn AMM) is a pre-built AI Agent focused on automating adverse media monitoring (AMM) or Negative News. Evan's tasks involve searching for and reviewing news about a person or company to determine if conducting business with them would incur a reputational risk or potential involvement with criminal activity. It is a vital part of the Know Your Customer (KYC) and Anti-Money Laundering (AML) processes for banks and other financial institutions.
Adverse Media Monitoring involves searching for negative news about a person or business. It can be a crucial part of Know Your Customer (KYC), and Anti-Money Laundering (AML) processes for banks.
Why our customers deal with Adverse Media Monitoring
Adverse Media scanning can also be considered a regulatory requirement. Regulatory bodies enforce banks to perform searches on companies, their top executives, and shareholders. Bank regulators demand the provision of reliable operation and security control over negative news processing. On the other hand, the current market requires a same-day screening service.
How Adverse Media Monitoring process is handled today
Research of negative news today is mostly manual and is therefore time-consuming and costly.
Major banks have dozens of analysts who scan news feeds (news aggregators). They review hundreds of articles throughout the analysis. Analysts make repeating efforts to go through articles and determine which of them are 'false positive'. Over 95% of articles are removed because of not being relevant to the initial search request.
Benefits of Adverse Media Monitoring automation
Automation of Negative News Screening brings you the following benefits:
- Screen larger amounts of information.
- Simplify and standardize the screening process.
- Reduce non-compliance risk by finding more relevant and complete adverse information.
- Reduce documentation time.
- Reduce manual efforts (gain of ~70%).
- Achieve real-time monitoring of events.
- Provide operational control and security.
- Improve employee satisfaction.
Implementation of Adverse Media Monitoring automation also brings the following transformational benefits:
- Consistent decision-making
- Faster response to customer requests
- Rapid scale-up of processing
- Better risk management through more instantaneous awareness of adverse news
Adverse Media sources
The AMM AI Digital Worker searches the following platforms for the negative news:
- Traditional news sources and media
- Social media and internet forums
- Regulatory filings and databases (for example, SEC and FINRA disciplinary actions)
- Blogs and other non-traditional media
- International Organization Databases (for example, the World Bank’s Stolen Asset Recovery DB)
Business Process overview
A “Business Process” represents the core workflow and design of any automation use case on WorkFusion’s Intelligent Automation Cloud platform, including Adverse Media Monitoring. The graphic below demonstrates the Business Process for the Adverse Media Monitoring skill.
In Adverse Media Monitoring, WorkFusion serves as a “maker” by gathering news content, processing the text, identifying articles with the highest risk materiality on the searched entity, and producing a detailed audit trail. Financial institution users (most often compliance analysts or account onboarding teams) become “checkers” by reviewing the skill's output. Through the Business Process detailed below, WorkFusion eliminates time spent on gathering data and reviewing false positives. The AI Agent enables financial institution users to prioritize their efforts on news results with meaningful risk events on the searched entity.
A false positive in this context refers to an article that does not include any material adverse news or an article that contains adverse information, but which is not focused on the searched entity.
The following numbers correspond to the workflow graphic above.
Negative news input. For each search request (for example, ultimate beneficial owner on a new corporate account application), WorkFusion fetches news articles by performing searches in pre-defined news sources and scraping text and metadata (for example, title, date, URL) of articles using APIs, if available, or robotic process automation (RPA).
Does news exist for entity? If no news articles are found for a searched name, the workflow is routed to step 5, in which an audit trail report is produced for each search. The audit trail report will note that no search results were available. No input or additional review from the user is required in the default workflow.
Negative news processing. WorkFusion processes the unstructured text of each article to identify the risk materiality. The machine learning approach, architecture, and logic used to conduct this analysis are detailed in greater specificity below. WorkFusion performs the following processing:
Identification of risk factors. Identification of high-risk countries, including sanctioned jurisdictions, and the relationship to the searched entity. This entity-relationship analysis identifies both positive (linked) and negative (not connected) associations to the searched entity.
Categorization of articles. Sentiment Analysis, a subset of Natural Language Processing, determines whether the article contains financial crime compliance-related adverse media. The model takes the following factors into account: the searched entity's relation to the article (that means are they the focal entity, are they quoted, and so on), searched entity relation to risk factors, and presence and relationship between keywords.
Identification of negative keywords. Identification of keywords and keyword phrases that contributed to WorkFusion’s categorization of the article.
Negative news review. After fetching, analyzing, and prioritizing articles, WorkFusion hands off the further verification to users within Workspace, where a user interface optimized for experience and speed of study presents the findings. Users review prioritized news articles and risk factors and negative keywords identified by WorkFusion to make a final determination on whether there is a material risk that requires escalation for the searched entity or an unrelated non-searched entity. The user interface includes:
Ranking. Based on the machine learning model's confidence scores of each article, the articles are stack-ranked from highest to lowest risk and color-coded to indicate which articles must be reviewed.
Explainer. Risk factors and keywords are highlighted in the article text. Users can add, remove, or modify these within the user interface. The keywords can also be used to conduct “scroll to the keyword” searching.
Summary. The most important parts of the article are displayed to the Workspace user, where the highest risk factors or keywords are found. You can view the full article text by selecting the article in the left column.
Negative news output. A detailed audit trail is created as a standardized PDF format containing the analyst decision on article materiality, cited articles, a confidence score for each article, static URL, keyword citations, and full article text.
Output analytics. Audit trail data is prepared for visualization in the WorkFusion Analytics dashboard.
Workspace is delivered “out of the box” within this skill. Financial institutions have embedded results within SharePoint, vendor workflow systems, and proprietary tools. This component would not apply in those cases.