This is a sample knowledge graph to demonstrate appliability in Capital Markets Compliance.
The Scenario (fictitious):
An institutional trader (Trader A) at Bank A has a meeting with executives at his public client Company X in Canada. She learns that they are planning to split into three companies. Likely Company X would engage Bank A to work on the deal – its shares have not yet been put on the Restricted List. Company X owns public Company Y in Mexico. One of the split Company X entities would canibalize Company Y. The trader has an acquaintance at an American broker (Broker B) who covers Company Y (i.e. market research). She makes a call to him and hints at Company X’s upcoming plans. Broker B sells Company Y short (buys an option), countervening Broker B’s compliance rules. Additionally Trader A has shared insider information, also braking Bank A’s compliance rules. The telecom companies analyze the call dialog to establish intent.
This small graph was built with Stardog Designer on Stardog’s Cloud. The data (i.e, class instances recorded as “Turtle” – Terse RDF Triple Language) were manually created and uploaded through Stardog Studio. The diagrams and videos are screen shots and recordings from Stardog Explorer.
See the screen recording here:
In the secreenshot below (from Stardog Explorer), start with the an options trade (at the top) that has taken place (lime), A compliance team would be responsible for examining trades against Securities laws (teal) and internal compliance rules (pale yellow). Policy 7.1 is a Retricted shares list. Clearly only a small selection of securities laws and compliance rules are shown here. In the lower left, you can see which compliance rules each financial service firm has internally established. In the scenario described above, you would expectthe trade to fail some laws and rules, as shown by the relationships outward from the trade. These would have been determined by the knowledge graph’s inference engine.
In the screenshot below (from Stardog Explorer), start with the cellphone call (light blue) between Victor and Jane (red). Presumably the call is digitized and recorded by the telecom company for each financial services firm (Audio for 2025-01-21 pale yellow). Either the telecom company (Bell, teal) would analyze the audio for keywords or the Banks/Brokers would do this in house. The analysis should result in a report that identifies calls needing human intervention/confirmation. You can see the Report for 2025-01-21 (beige) identifies the keyword Buy (purple); this relationship likely would have been created by Stardog’s inference engine.
