In early 2020, in a compliance department, an analyst typed the words "Hin Leong" into a search bar and saw nothing alarming. The name seemed clear, and there was nothing to suggest the relationship carried any risk.
The analyst was partially right but ultimately wrong. The company collapsed without appearing on any sanctions or PEP list. But the case of Hin Leong was not invisible. Before the bankruptcy filing, before any official designation, the signals were already in the public domain. Bloomberg had reported on the company's financial difficulties as early as April 10. The Business Times had covered it. The signals were public, available, and indexable.
For institutions with adverse media monitoring in place, those reports would have triggered an alert before the April 17 bankruptcy filing. That is a window to review the relationship, reduce exposure, or escalate internally. For institutions relying only on official lists, that window did not exist.
The Hin Leong case was not an anomaly. It reflected a gap that regulators had already identified and were actively working to close.
At the global level, FATF recommends that institutions deploy verifiable adverse media searches as part of enhanced due diligence, treating negative news coverage as a signal of risk that may require additional scrutiny. In the EU, the requirement was progressively strengthened across successive directives. The 4th Anti-Money Laundering Directive, which came into effect in June 2017, introduced the first formal requirement for firms to screen against open source media as part of enhanced due diligence. The 5th AMLD expanded that requirement to more sectors and pushed institutions toward automation. The 6th AMLD extended the list of predicate offences linked to money laundering to include cybercrime and environmental crime.
The regulatory expectation, in other words, has only grown more demanding with time. The operational reality at most institutions has not kept pace.
What adverse media screening is supposed to do versus what it is actually doing
The regulatory expectation clearly demands continuous, structured monitoring of public sources, categorized by risk type, integrated into the customer due diligence process. The operational reality at most institutions is considerably different. Adverse media screening, the practice of systematically monitoring public news sources for negative coverage linked to financial crime, reputational risk, or regulatory violations, is where that gap shows up most clearly.
A 2023 survey of 205 compliance professionals found that 84% of firms remain reliant on manual screening processes, despite automated tools being widely available and 5AMLD explicitly pushing institutions toward automation in screening of media outlets. Manual screening using internet search engines is time-consuming, inconsistent, and prone to human error. It is also the method most firms are still using.
In practice, compliance teams waste 100 or more hours annually reviewing false positives, delaying decision-making and increasing operational costs. The problem is not expected to ease: 40% of firms believe that the challenge of dealing with false positive alerts will increase, and almost 50% believe keeping pace with regulatory change will become harder.
The root cause is structural. Keyword-based approaches search using predefined terms associated with financial crime. While cost-efficient, they generate high false positive rates due to the signal-to-noise problem, where legitimate mentions are conflated with genuinely adverse information. Hybrid approaches combining keyword searches with human expert review achieve better precision but at significant cost and with limited scalability. The operational cost extends beyond time. A survey of 600 financial crime decision-makers found that 53% cited removing repetitive, no-value alert remediation as the primary benefit they would seek in a new solution. That is not a feature preference. It is a description of where analyst capacity is currently going.
The question is not whether adverse media screening should be part of a compliance program. Regulators have settled that. The question is whether the tools institutions are using are actually capable of delivering what regulators expect, and what the consequences are when they are not.
What changes when adverse media screening is built around risk, not keywords
Advanced Adverse Media is Sanction Scanner's answer to that gap, built into the AI-powered risk intelligence FUSION platform and designed to replace classification work that currently falls on the analyst.
For the analyst, the most immediate difference is what they see when they open a case. Instead of a raw list of articles sorted by publication date, they see a structured risk profile. Results already classified and severity already scored. The decision-making work starts at the top of the queue, not after an hour of sorting.
That shift matters because the volume problem is real and growing. Approximately two million news articles are published globally every day. Without classification at the point of ingestion, compliance teams are not screening media. They are searching it, which is a fundamentally different and far more resource-intensive task.
AI classification across six risk categories
Advanced Adverse Media classifies every result into one of six dedicated risk categories: Financial Crime, Terrorism and Sanctions, Violent and Organized Crime, Legal and Regulatory, General Adverse Media, and Cyber and Data Security. Each result is then scored High, Medium, or Low.
The value of category-aligned classification is not just operational. It is regulatory. When categorization covers only general news topics, it loses its AML and CFT risk-based perspective, which means firms may fail to meet requirements under FATF and the EU AMLDs because they cannot adequately choose categories according to relevant risks.
The six categories cover the core offense types across FATF guidance and the 6AMLD predicate offence list, with particular strength in financial crime, terrorism, organized crime, and cybercrime.
Coverage that reflects where financial crime actually surfaces
Local reporting frequently breaks stories days or weeks before international media notice them, if they notice them at all. Across emerging markets, corruption investigations and financial misconduct allegations frequently appear first in regional business press, specialist investigative publications, or local court notices.
Advanced Adverse Media draws from globally indexed news media in real time. Coverage spans nine languages: English, Turkish, German, Spanish, Italian, French, Arabic, Portuguese, and Russian. Search runs natively in each language, meaning a name searched in Arabic returns results from Arabic-language media directly, not translated approximations of English coverage. Coverage that stops at English-language headlines is not global coverage. It is partial coverage with the appearance of completeness.
Integration inside Fusion
Adverse media results on their own are useful. Adverse media results viewed alongside sanctions data, PEP status, transaction patterns, KYB findings, and customer risk scores tell a different story entirely.
Advanced Adverse Media sits inside the FUSION platform. A compliance analyst reviewing a customer does not need to cross-reference a separate adverse media tool against a separate sanctions system. Every risk signal for that customer is visible in one place, in one workflow. Consolidating screening and monitoring into a single, integrated solution optimizes resources and improves efficiency through automation.
This integration also closes a specific gap that siloed tools create. A sanctions hit in isolation may not be enough to escalate a case. But a sanctions near-match combined with an adverse media result categorized as Financial Crime, combined with an unusual transaction pattern, creates a materially stronger case for enhanced due diligence or exit. That combination is only visible when the data lives in the same environment.
See Advanced Adverse Media Screening in action, try our FUSION platform
