Creating your own screening lists
Sanction and PEP lists are pre-integrated, but the custom lists an institution actually needs do not: industry-specific watchlists, official gazette decisions, and internal blacklists. TruvaLI builds these lists in three ways: by defining an official source address, setting up a feed, or uploading a document. In each case, names, aliases, and identity details are extracted and turned into screening-ready records.
Why pre-integrated lists are not enough
Sanction and PEP lists come from international sources and are identical for everyone. The lists an institution actually needs often fall outside these databases.
For a betting operator, this means officials who can influence game outcomes, for a payment institution, high-risk merchants, for a bank, its own internal blacklist, and for a crypto platform, specific wallet addresses. None of these are available in commercial databases because none of them are public or generic.
Maintaining these lists manually is impractical. When a decision is published in the official gazette or a sports federation announces a squad, the person responsible for manual updates is often occupied with other tasks.
Method one: official source address
When defining a source, you provide its address, language, and country. Three screening parameters are then added:
- Link pattern to follow. The links to be monitored are restricted by a pattern, meaning the crawl targets only the relevant section rather than the entire website.
- Depth and page limit. This defines how many levels deep the crawl should go and the maximum number of pages to read. Crawling is not unlimited.
- Screening frequency. This determines how often, in hours, the source is re-scanned.
A subject code and an analysis prompt preset are linked to the source. Content from the source is processed using this prompt and converted into records. You can see exactly when each source was last synchronized.
Method two: media feed
An adverse media feed is defined by specifying keywords, language, country, how many days to look back, and the maximum number of news articles to retrieve per page.
Articles from the feed are first queued, then converted into records after processing. The system logs when the feed was last polled.
Method three: document upload
An official gazette, court ruling, regulatory circular, or the institution's own list is uploaded as a file. An analysis prompt preset is selected for the upload, and the document is processed. You can monitor the status and completion time of the operation. The raw response is also stored, allowing you to audit exactly what the extraction process read.
Analysis prompt presets
Extraction is not bound to a fixed template. An analysis prompt preset has a name and a template. Institutions define presets based on their specific needs and link them to sources, feeds, or documents.
Extracting names from an official gazette decision is not the same as extracting names from a sports federation squad list: the preset manages this distinction.
What the record contains
Each extracted record contains more than just a single name: first name, last name, corporate title, aliases, gender, date of birth, birth year, and city. The record is linked to a subject and a sanction authority, with the authority's country code stored separately.
Having the birth year as a separate field from the full date of birth is intentional: official sources often only provide the year, and requiring a full date would make the record unusable.
A record can be flagged as either a sanction record or an internal blacklist record. Both enter the same screening flow, but their sources remain clearly distinguishable.
Integration into the screening flow
The created list does not sit in isolation. It is evaluated in the same flow as pre-integrated lists during customer onboarding, post-onboarding monitoring, and within the rule engine. In crypto contexts, wallet addresses are also screened as entities.
Benefits to the institution
- The institution's custom list covers risks that fall outside commercially available databases.
- When an official decision is published, the list is not updated manually; the source is automatically re-scanned.
- What the extraction process reads is stored, meaning the list is not a black box.
- Because records include aliases and partial birth details, matching quality does not rely solely on a single name.
Common questions
- How deep does the crawl go when we provide a website address?
- The link pattern to follow, depth, and maximum page count are defined separately. The crawl targets only the section specified by the pattern rather than the entire site, and it is not unlimited.
- Can we see what was extracted from an uploaded document?
- Yes. You can view the status and completion time of the process, and the raw response is stored so you can audit exactly what the extraction process read.
- Can we define our own extraction templates?
- Yes. Analysis prompt presets are structures with a name and a template, which are linked to sources, feeds, and documents. This distinction is necessary because extracting names from an official gazette decision is different from extracting them from a squad list.
- Does the record only store names?
- No. It stores the first name, last name, corporate title, aliases, gender, date of birth, birth year, and city. The birth year is a separate field because official sources often only provide the year.
- Do internal blacklists and sanction records get mixed up?
- No. A record can be flagged as either, and both enter the same screening flow, but the source of the record and its associated authority remain clearly distinguishable.