Skip to main content Link Search Menu Expand Document (external link)

Artifical Intelligence

  1. Overview
  2. Models
  3. Trace Model Type
    1. Standard
    2. Pre-Built
    3. On-Demand
  4. Fields
  5. Model Details
    1. Risk Detection
    2. Language Identification
    3. Spam Detection
    4. AI Extracted Text Cleansing
  6. Additional Information

Overview

Trace provides out-of-the-box Artifical Intelligence (AI) to improve the user experience, reduce review time, pinpoint risk.

Models

Trace Model Category Trace Model Type Relativity Model Model Type Model Description
Irrelevant Content Detection Standard AI Extracted Text Cleansing NLP Identifies non-authored or duplicative content, email headers, email signatures, and confidentiality footers
Irrelevant Content Detection Standard Spam Detection Classification Spam is defined as irrelevant communications that could not contain misconduct
Irrelevant Content Detection Pre-Built Newsletter Detection Classification Detect widely disseminated and recurring messages that summarize world and market news (“Newsletters”) that do not need to be reviewed, to reduce false positives
Irrelevant Content Detection Pre-Built Financial Research Report Detection Classification Detect widely disseminated documents prepared by investment research firms that describe a company and often provide a particular recommendation (“Research Reports”) that do not need to be reviewed, to reduce false positives
Reviewer Understanding Standard Language Identification NLP Identifies the primary and other languages in a communication
Reviewer Understanding Pre-Built Audio Trascription NLP Transcibes audio communciations into text
Risk Detection Pre-Built Change of Venue Classification Detect an attempt to avoid discovery by changing the communication venue
Risk Detection Pre-Built Rumors And Speculation Classification Detect the distribution or discussion of unverified and doubtfully true information
Risk Detection Pre-Built Collaborative Discussions Classification Detect employees working together to prevent the discovery of misconduct and sharing of client identifying data
Risk Detection Pre-Built Boasting Financial Classification Detect excessively proud and self-satisfied talk by a trader of their achievements
Risk Detection Pre-Built General Tipping Classification Detect the sharing of potential material nonpublic information regarding company & financial information
Risk Detection Pre-Built Prearranged Trading Classification Detect traders agreeing on trade price in advance of execution
Risk Detection Pre-Built Price and Benchmark Fixing Classification Detect agreements between traders to collectively manipulate the price of a financial product or a benchmark to adjust market conditions
Risk Detection Pre-Built Front Running Classification Detect traders attempting to buy or sell financial products ahead of other orders to lock in a better price
Risk Detection Pre-Built Mirror Trading Classification Detect simultaneous buy and sell trades of the same stock on different exchanges by the same entity, as a way to move money between jurisdictions
Risk Detection Pre-Built Forex Collusion - Trade Fixing Classification Detect agreements between traders to collectively manipulate benchmark foreign exchange rates (FX) to artificial levels

Trace Model Type

Standard

AI functionality enabled for all customers that is built by Trace and requires no customer training. Standard models are focused on understanding that doesn’t change across customers or industries

Pre-Built

AI functionality avaliable to all customers that is built by Trace and can be improved by a customer. These models support validation and are focused on common compliance requirements.

On-Demand

AI functionality that a customer can define and build. These are focused on customer specific requirements.

Fields

Document fields are created for each enrichment method. These document fields can be used as part of the Omit from Alert Rules functionality or Rules. See more information on AI fields here.

Model Details

Below is supplemental information to help understand and implement the models above.

Risk Detection

All Risk Detection information can be found here

Language Identification

For infromation on enablment can be found here

Spam Detection

For infromation on enablment can be found here

AI Extracted Text Cleansing

For infromation on enablment can be found here

Additional Information

For more information on how Trace’ AI models are enabled, please see the Admin Guide