Skip to content
NegativeNewsSearchA DemandFlow product

Adverse media screening · AML / KYC / KYB

Adverse media screening that knows where it cannot see.

General web search returns fewer and fewer neutral and primary sources for negative news research. NegativeNewsSearch combines country-aware search, a 36-month GDELT archive, deterministically resolved name variants and matching against the EU, OFAC and US Consolidated Screening lists into one structured report — per name, in minutes.

A laptop on a glass table showing an analysis view with charts.
risk dimensions, scored 0–3 per article with a stated rationale
4risk dimensions, scored 0–3 per article with a stated rationale
months of GDELT archive, filtered to 13 negative-news themes
36months of GDELT archive, filtered to 13 negative-news themes
sanctions sources: the EU consolidated list, OFAC SDN and the US Consolidated Screening List
3sanctions sources: the EU consolidated list, OFAC SDN and the US Consolidated Screening List
HTTP header — that is the entire integration
1HTTP header — that is the entire integration

The problem

You know how this goes.

Adverse media research rarely fails because nobody searched. It fails on where the search ran, which spelling it used — and what can still be evidenced two years later.

  1. Page one is search engine optimisation.

    The name goes into a web search. What sits at the top sits there because someone worked to put it there. The local newspaper report from the country where the company actually operates is on page seven — or simply absent from the index of the domain you queried.

  2. The hit was filed under a different name.

    The subsidiary trades under a different name from the parent, the transliteration differs from the original, the short form used in the market differs from the register entry. A search across exactly one spelling is a search with a foregone conclusion.

  3. No hit is not a result.

    In a country with severely restricted press freedom, local search finds nothing. That is not exoneration, it is a missing source. Yet the file ends up reading exactly as it would for a genuinely clean subject: "nothing found".

  4. And then the audit arrives.

    A folder of screenshots, a handful of links, a file note written from memory. The question is never "did you search?" — it is "how, where, when, and why did you assess it that way?".

A newspaper stand at night, many titles side by side, blurred street lights behind.

What NegativeNewsSearch does

One name, one country, one report.

You submit a name and a country. What comes back is not a result list somebody still has to work through, but an evidenced and scored assessment.

Searches wider than a single search box

Live search on the right country domain, a 36-month archive drawn from the GDELT Global Knowledge Graph, and sanctions list matching all run within the same job. Results are merged and deduplicated before anything is analysed.

Actually reads the articles

Articles found are fetched in full text — with a browser fallback for pages that refuse a plain request — and assessed individually. The judgement is made on the article body, not on the two-line preview from the results page.

Scores in a structure instead of narrating

Per article: a fact-only summary, an explicit statement of how it relates to the subject, and four risk scores from 0 to 3 — each with a rationale that has to come from the text. Above that, an overall assessment across all articles.

Documents itself

Every statement carries a source and a publication date. Articles that could only be retrieved in part are flagged as truncated rather than quietly papered over. The report is the audit document, not the raw material for one.

The run

What happens between request and report.

Seven steps, exactly one of which you have to build.

  1. 1POST /tasks

    Request

    A name and a country. One HTTP call with one extra header — that is the whole of the integration work on your side.

  2. 2Aliases

    Resolve name variants

    The name is resolved against a curated set of trading names, short forms and transliterations, and submitted as an OR block — deterministically, from a maintained list, not guessed by a language model.

  3. 3RSF rank

    Choose the search lanes

    The country determines search domain, language and region. Where the country falls into the range of severely restricted press freedom, NegativeNewsSearch opens an additional international lane.

  4. 4GDELT

    Query the archive

    36 months of Global Knowledge Graph, pre-filtered to 13 negative-news themes — from money laundering through sanctions and organised crime to tax offences.

  5. 5EU · OFAC · CSL

    Match against sanctions

    Fuzzy name matching against the EU consolidated list, the OFAC SDN list and the US Consolidated Screening List. Each hit carries its programme, source, data vintage and similarity score.

  6. 6Full text

    Analyse the articles

    Fetch, extract, assess individually. Four risk scores with rationales — and, before that, a finding on whether the article concerns the subject at all or merely mentions it in passing.

  7. 7Report

    Consolidate

    One overall assessment across the individual analyses, derived from nothing else. Retrieve it by polling, by live stream, or by webhook.

How it works, in detail

Output

This is what you get back.

Not prose you have to mine for the assessment. A fixed structure you can file as it stands.

Subject

Musterhandel International S.A.

Country

LY — Libya · international lane active

Summary

Prosecutors have been investigating two former board members of the company since February on suspicion of commercial bribery. The proceedings concern payments connected to the award of an infrastructure contract, said to have been routed through an intermediary company. The company rejects the allegations.

Connection to the searched organisation/person

The searched organisation is the direct subject of the article. The named individuals are expressly described in the text as former board members of the company.

Risk assessment

  • Money laundering / terrorist financing

    The article describes payment flows through an intermediary company but raises no money-laundering allegation.

  • Sanctions / embargoes

    No indications in the article text.

  • Fraud / corruption

    Ongoing criminal investigation on suspicion of bribery. No legally binding finding stated in the article.

  • Tax

    No indications in the article text.

Source: beispiel-zeitung.example · business desk

Published: 14 March 2026

This article may have been truncated by a paywall and could not be read in full.

Found but not assessed

14 assessed · 4 could not be read · 9 not reached

The card above is one of the fourteen assessed articles. These four were found by the same screening and judged relevant, but could not be read automatically — so they are handed over as links rather than dropped. Their absence from the scores is not a finding of "no risk": nothing is known about their content, and the overall score does not account for them.

  • Access blocked — paywall or access restriction

    Investigative feature on the intermediary company

    https://recherche-zeitung.example/dossier/zwischengesellschaft

  • Access blocked — paywall or access restriction

    Interview with the former compliance officer

    https://wirtschaftswoche.example/interview/compliance

  • No readable article text

    Prosecutors’ press release (PDF)

    https://staatsanwaltschaft.example/presse/2026-03-14.pdf

  • Not reachable

    https://regionalblatt.example/2026/03/vergabe-infrastruktur

Sample output using an invented company, an invented set of facts and invented counts — the unreadable sources listed above are invented too. It shows the format, not a real finding about a real business. The proportions are realistic: on live queries a meaningful share of relevant hits cannot be read automatically, and paywalls fall hardest on investigative reporting.

What sets us apart

Three things no other vendor does this way.

Press freedom is a search parameter

Where the local press cannot report freely, local search is not a reliable source. NegativeNewsSearch identifies those countries by their RSF rank and searches internationally as well — automatically, not on request.

Unlinkable — not by promise, by mathematics

Which names you screen is itself confidential information. Access runs on blind-signed tokens per RFC 9474: the server cannot attribute a redeemed query to the customer the token was issued to.

Name variants without guesswork

The OR expansion comes from a curated set, not from a language model. The same input produces the same search — today, next quarter, and in the audit meeting.

See all the differences

We publish a page about the product’s limits.

Adverse media screening is decision support, not a decision. What NegativeNewsSearch cannot do, does not cover and does not replace is spelled out on its own page — before the first conversation, not in the middle of due diligence.

Read pros & cons

Test it against your own cases.

Give us a handful of names where you already know what should have been found. That is the only test that tells you anything.