LSEG Risk Intelligence, the risk intelligence business of LSEG, has published global research suggesting that while compliance leaders almost universally trust their screening data, far fewer can readily prove how their decisions were reached.
The survey questioned 600 senior risk and compliance decision-makers across EMEA, APAC and the United States. Some 94% of participants expressed confidence that their screening data is both complete and up to date. Yet under half, 48%, believe they could very easily hand a regulator a full audit trail on request, and 41% admit their organisation is not entirely ready for more intensive regulatory examination.
LSEG Risk Intelligence has labelled this shortfall the ‘decision gap’. The term describes the practical space between flagging a possible risk and holding the context, decision logic, ownership and evidence required to act on that risk and justify the action at a later date.
Manual checks remain a core part of how firms bridge this gap. Almost every respondent, 93%, cross-checks screening outcomes against a second source at least some of the time, and 69% do so frequently or always before acting. According to the research, this shows human review still carries significant weight in screening decisions, but it also means firms must devote substantial operational resources between receiving an alert and arriving at a decision they can defend.
The study also looked at data quality. Internal data strength and availability was rated highly by 83% of respondents, and external reference data by 84%. However, only 76% gave high marks to both. When asked what they hoped to gain from better data quality, explainability and matching systems, respondents placed more confident and consistent screening decisions level with stronger audit trails and regulatory evidence, each chosen by 39%.
On the subject of AI-supported screening, participants outlined the safeguards they would need before placing greater trust in automated outcomes. Trusted, high-quality data topped the list at 40%, followed by the ability to explain every decision to regulators at 37%, regular accuracy and performance testing at 36%, and keeping human review within the process at 35%. The research concludes that confidence in AI-assisted screening will rely on firms pairing suitable technology with reliable information, proven performance, explainability and human oversight.
LSEG group head of risk intelligence David Wilson said, “Compliance teams around the world are confident in the data they use, but confidence alone is not enough. Organisations need to connect trusted information, clear decision logic, human judgement, and accessible evidence so they can explain the actions they take.
“The opportunity is not simply to add more data or automation. It is to create a controlled route from a risk signal to a decision that remains explainable and defensible when it is examined months or years later.”
Copyright © 2026 RegTech Analyst
Copyright © 2026 RegTech Analyst
