Glossary

Compliance concepts

Adverse media

Negative coverage of a person or organisation in public news and media sources. Searched as a risk signal in its own right, not only as a supplement to a list match. Used on /solutions/adverse-media-search/

Alert

A match or a pattern that has crossed a configured threshold and now needs a human decision. Every module raises alerts into one queue. Used on /platform/alert-management/

Beneficial ownership

Who ultimately owns or controls an organisation, as distinct from who is named on its paperwork. Where sanctions exposure in a supply chain usually sits. Used on /solutions/third-party-due-diligence/

Disposition

The recorded outcome of an investigation: the decision, who made it, when, and any notes. What survives an examination. Used on /platform/alert-management/

False positive

An alert that investigation shows to be no risk. The cost of screening is mostly the review of these, which is why reducing them is treated as an outcome rather than a side effect. Used on /solutions/name-screening/

KPI and KRIKey performance and key risk indicators

The measures a risk profile is calculated from, recalculated continuously rather than on a review cycle. Used on /solutions/risk-based-profiling/

KYCKnow your customer

Establishing and maintaining who a customer is, at onboarding and continuously afterwards, so that a customer who becomes a match later is found without a manual re-run. Used on /solutions/name-screening/

PEPPolitically exposed person

Someone holding a prominent public function, or close to someone who does, whose position carries a higher risk of bribery or corruption and therefore a higher standard of diligence. Used on /solutions/name-screening/

Risk tier

The band a customer's risk score places them in, which determines the level of diligence applied to them. Used on /solutions/client-risk-scoring/

Risk-based approach

Directing effort in proportion to assessed risk rather than uniformly. Something a system does continuously, rather than something a policy asserts. Used on /solutions/risk-based-profiling/

Sanctions list

A published list of persons, entities, vessels or instruments subject to restrictive measures. Screened against continuously, because the lists change rather than the customer. Used on /solutions/transaction-filtering/

Whitelisting

Recording a confirmed false positive so the same match does not raise the same alert again. Used on /solutions/name-screening/

Detection techniques

Two engines named on the platform, Clustering and Modelling and Evolutionary Learning, are absent here on purpose. Both are marked in preparation on their own page, with no verified description to publish, and defining them here would invent what that page declines to state.

Batch and real time

Whether a check runs on a schedule over accumulated records, or in line as the record arrives. The difference between reversing a payment and blocking one. Used on /solutions/transaction-filtering/

Behavioural analysis

Behaviour assessed over time rather than per transaction. A baseline is built per customer, so a payment is judged against what that customer normally does instead of against a global threshold. Used on /platform/detection-engines/

Business rules engine

Where the detection logic lives. Rules are versioned artefacts: a new version is published when ready, previous versions are retained, and a rollback is a selection rather than a rebuild. Used on /platform/business-rules-design-and-engine/

Dictionary

Static reference data a rule looks values up in, maintained through the Rule Designer alongside the rules themselves. Used on /platform/business-rules-design-and-engine/

Entity matching

Resolving whether two records in datasets that share no identifier refer to the same person or organisation, using probabilistic matching and deduplication. Used on /platform/enhanced-ai-search/

Fuzzy string matching

Matching that tolerates difference, so a name survives a spelling, a transliteration or a reordering. Where inconsistent source data actually lives. Used on /platform/enhanced-ai-search/

Hierarchical clustering

Grouping records by similarity at successive levels of granularity, used to surface relationships that are not stated in the data. Used on /platform/enhanced-ai-search/

Instant payment scheme

A payment scheme that settles in seconds, so any screening check has to complete inside the scheme response time rather than in a batch window. Used on /solutions/transaction-filtering/

Latent Semantic Analysis

A technique for relating documents and terms by the contexts they share rather than by exact wording, used in scoring and in surfacing hidden relationships. Used on /platform/enhanced-ai-search/

Natural language processing

Reading unstructured sources such as news, media and public registers, scoring each document for risk and sentiment from keyword frequency, sentiment analysis and topic relevance. Multiple languages are supported. Used on /platform/detection-engines/

Pattern recognition

Statistical and behavioural models applied to transaction and customer data to find patterns a single rule would not describe. It works alongside the rules engine rather than instead of it. Used on /platform/detection-engines/

Profiling and scoring

A risk score and risk tier per customer, recalculated as new information arrives from screening, monitoring and the customer record. The rules and events that moved a score are retained alongside it. Used on /platform/detection-engines/

Ruleset

The collection of rules a module evaluates against. One a compliance team maintains ages at the rate the risk does; one only a vendor can change ages at the vendor's release cadence. Used on /solutions/transaction-monitoring/