Matching is the whole problem
The screening core merges optimised search algorithms with fuzzy matching, entity disambiguation and predictive search, so a name is resolved rather than merely compared. Screening runs against multiple data sources in parallel, and every match carries a similarity score.
What the screening core does
A list check is easy. Deciding whether two names are the same person is the work, and it is where false positives come from.
Fuzzy string matching
Match
Identifies partial or approximate matches, capturing near-misses and inconsistencies that an exact comparison drops, and returns a similarity score with each one.
Optimised entity matching
Resolve
Matches entities such as names, transactions and assets across disparate datasets, using techniques such as probabilistic matching and deduplication algorithms.
Semantic search
Read
Applies vector-based matching and natural language processing to find contextually relevant results, beyond what a keyword search returns.
Risk scoring and alerts
Score
Risk scoring runs on the results, so a suspicious entity or behaviour is flagged and the critical actions are prioritised.
Real-time processing
Now
Processes, analyses and retrieves results as the query runs, which is what makes a check at onboarding or at payment possible.
Multiple lists at once
Parallel
Checks in parallel against multiple data sources such as sanctions lists, PEP data, internal blacklists and whitelists, via a standard list format.
Where it is used
The screening core is shared. These solution pages run on it.
Solution
AI search with fuzzy string matching against global watchlists, individually and in fully automated batch runs.
Solution
Assets cross-referenced against international sanctions and embargo lists, with a similarity score on every result and alerting on defined thresholds.
Solution
Fuzzy matching across unstructured and structured sources, with similarity scoring, so a mention is tied to the right entity.
Solution
Counterparties screened against the same sanctions lists and watchlists as customers.
Solution
Search across structured and unstructured data for identity due diligence, beside the behavioural signals.
Common questions
No. Alongside fuzzy string matching the core applies semantic search, using vector-based matching and natural language processing to find contextually relevant results beyond a keyword approach. Advanced pattern recognition and risk scoring run on the same results, so a hit arrives already prioritised.
Sanctions lists, PEP data, internal blacklists and whitelists, through a standard list format. Which commercial providers you use is your subscription rather than ours.
Both. Individual name screening runs online and batch screening runs fully automated, against databases that are continuously updated and monitored.
The score comes from the matching and is shown with the match. Where your thresholds should sit is a calibration question for a working session, not a number a page should quote.
Your analysts do, with help. The live material describes an automated four-eye control in which a large language model assists and automates the review process, significantly reducing resource demand. Whether your policy allows a model in that seat is a question for your own second line.
Yes. The live material says screening is not limited to names and also covers companies, vessels, securities and other entities.
Name Screening Intelligence · DetectX®
See a match scored and explained.
Book a working session on your own list set. Bring your questions about matching, thresholds, white lists and batch scheduling.
- Fuzzy matching
- Partial and approximate matches captured, each with a similarity score.
- Parallel lists
- Sanctions, PEP data, blacklists and whitelists checked together.
- Beyond names
- Companies, vessels, securities and other entities.