Adverse media screening methodology
Adverse media screening on Stonewake reviews public news and open sources for serious misconduct linked to organisations and countries, with quote-verified findings and analyst confirmation. Draft verdicts are offered for human review. The product does not perform sanctions or PEP list matching and does not issue credit scores about natural persons.
Institutional context for adverse media screening
Bank customer due diligence programmes already treat negative news as a risk-assessment input. The European Banking Authority's ML/TF Risk Factors Guidelines ask firms whether there are adverse media reports or other relevant sources about the customer or the ultimate beneficial owner, and whether those reports are reliable and credible. Credibility depends on the quality and independence of the publisher and how persistently a story is reported, among other factors. A conviction is not required for an allegation to remain relevant.
FATF standards frame the national AML/CFT systems within which banks operate KYC and KYB processes. Adverse media screening sits beside identity, ownership, and sanctions screening. It is broader and more judgmental than list matching, and it generates noise as well as signal.
Scope: organisations and countries
At product capability level, Stonewake adverse media screening focuses on organisations and countries. Findings quote-verify allegations against stable public sources under the citations and evidence standard. Screening covers multiple allegation categories over a multi-year lookback window, described on the site as nine categories across a five-year window.
Results are presented as draft verdicts in the Red, Amber, Green, or Nothing found family for analyst confirmation. Confirmed dispositions feed an append-only audit ledger so a later reviewer can see who confirmed or dismissed a finding. Exact engineering internals, vendor lists, and crawling recipes are outside the scope of this methodology page.
What a finding includes
A usable adverse media finding on Stonewake typically exposes:
- the screened organisation or country subject
- the allegation category in product taxonomy
- a verbatim quote from a cited public page
- an evidence label consistent with Verified or Reported discipline
- a draft verdict awaiting analyst confirmation
Proximity analysis remains a desk task: whether the named entity is the bank's customer, a controller, a sponsor, an EPC contractor, or an unrelated namesake. Export finance, project finance, and CRE files often involve layered ownership and cross-border sponsors, which is why glossary concepts such as beneficial ownership and UBO identification remain part of the surrounding diligence stack.
Boundaries with sanctions, PEP, and credit decisions
Stonewake adverse media screening methodology is explicit about non-goals:
- It is not official sanctions list screening.
- It is not a politically exposed person determination product.
- It is not automated credit scoring.
- Draft verdicts require human confirmation before they are treated as desk disposition.
Binding prohibitions remain the domain of sanctions programmes and a bank's designated screening tools. Adverse media may surface risk narratives before a party appears on a list, but only list-based controls create listing consequences.
Fit to EF, PF, and CRE workflows
On export finance desks, adverse media often attaches to buyers, guarantors, exporters, and agent banks around buyer credit and export credit agency supported structures. On project finance desks, sponsors, offtakers, and contractors sit inside long-tenor files where reputational and financial-crime issues can reopen conditions precedent. On CRE desks, borrower groups and cross-border owners raise similar open-source questions beside property metrics.
Commercial intent surfaces that reuse this methodology include counterparty adverse media for banks and jurisdiction context in country risk for banks. Related hubs: sanctions and AML trade finance, export finance, and project finance.
Evidence quality and analyst confirmation
Because adverse media is judgmental, the methodology privileges quote verification and human disposition over silent automation. An analyst can confirm, amend context, or dismiss a draft. The product posture is human in the loop: scores and verdicts are drafts, not unsupervised decisions producing legal effects by themselves.
Source quality still requires desk standards. Tabloid repetition, outdated stories, and mistaken identity are known failure modes. Documented disposition criteria inside a bank's financial-crime framework remain authoritative. Stonewake supplies cited research artefacts; the bank owns the risk decision.
Compliance and residency constraints that surround screening data are summarised in data residency and compliance. Country-level composites that may incorporate verified adverse news are described in country risk composite.
Category coverage without teaching evasion
The site describes multiple allegation categories over a five-year window. This methodology names that capability without publishing a how-to for gaming sources or suppressing hits. Mapping product categories to a bank's internal taxonomy for bribery, fraud, environmental crime, sanctions evasion narratives, and related financial-crime themes is a compliance design task owned by the institution.
Nothing-found draft verdicts are not certificates of cleanliness. They mean the screening run did not surface quote-verified material meeting product thresholds in the lookback window. Customer risk assessment judgment still applies, particularly for thin-file jurisdictions or low-visibility corporate groups.
Record-keeping for second line and audit
Analyst confirmation with an append-only audit ledger supports later reconstruction of who accepted or dismissed a finding. That matters when internal audit or a supervisor asks how open-source hits were handled on a buyer-credit or project file. The methodology does not prescribe a bank's retention schedule; it supplies product behaviour that can align with regulated record-keeping when configured under a bank's own policies.
Screening outputs pair with identity and ownership work already documented under KYC and KYB procedures. Adverse media without a clear link to the customer or controller wastes committee time. Clear linkage with weak sources still requires credibility analysis under EBA-style reliability tests.