What sets Global Relay’s AI-enabled surveillance approach apart?

—

What sets Global Relay’s AI-enabled surveillance approach apart?

Generative AI offers compliance teams risk identification abilities that completely transform surveillance operations.Contents

—
About Global Relay

Generative AI offers compliance teams risk identification abilities that completely transform surveillance operations.

Levelling up from lexicon-based systems, Global Relay utilizes a Large Language Model (LLM) to review whole conversations rather than isolated sentences, and determine the sentiment behind messages. This allows for more comprehensive risk justification and eliminates high volumes of false positives.

Global Relay’s open-source LLM has profound built-in knowledge extending far beyond financial services. While other AI-enabled surveillance systems rely on smaller deep learning models targeting only finance-specific risk, Global Relay’s LLM stands apart. Monumental in scale, it draws a significant level of knowledge and sophistication from endless data points to analyze communications and determine if risks are genuine, eliminating the bias and tunnel vision that can come from narrowing knowledge to a single industry.

Refined by a team of over 300 experts, our LLM has been steered to perceive risk areas existing within finance, commodities, and regulatory compliance domains. With this knowledge, it can execute an accurate risk analysis of your business communications.

Comprehensive risk coverage

Our LLM is aligned to 11 key risk categories encompassing over 130 risk indicators – capturing the conduct and compliance threats relevant to your business.

Our risk categories span critical areas, including client interests, financial crime, market abuse, collusion, compliance oversight avoidance, inducements, and more. Each category is further refined into over 40 sub-categories, addressing specific concerns like fair and balanced advice, excessive trading, bribery, money laundering, unapproved communication channels, price fixing, and insider trading. Due to its extensive knowledge, minor adjustments to prompts allow our LLM to maintain peak performance as risks emerge or evolve.

Secure and private hosting

Global Relay’s LLM is hosted within our privately owned data center, meaning your data remains securely within the Global Relay ecosystem for analysis – we will never share it with third parties for processing.

We refer to in-house data scientists and subject matter experts to finetune our system using synthetic examples that reflect real-world risks – ensuring that clients’ data privacy is never compromised. Having a private data center allows us to process more data at the same cost a third party would charge for processing significantly less, allowing us to achieve optimal performance.

Upgrading over retraining

Our AI model does not require lengthy retraining, due to its considerable size and capability.

With simple prompt adjustments, it quickly adapts to emerging and evolving risks, including those specific to your business. To ensure you are always utilizing the most advanced AI for accurate communications monitoring, we upgrade and replace our model in alignment with the continuous innovation in AI.

Explainability and transparency

To address explainability concerns that can accompany AI-generated risk alerts, we maintain a comprehensive risk governance framework that adheres to global AI governance principles and model risk management guidance. This governance illustrates how our LLM is managed and trained, and how information is documented.

Human interpretation of AI recommendations is crucial for a successful AI-enabled surveillance workflow. To ensure complete transparency and accuracy, our LLM simultaneously generates a clear explanation each time it flags a message as a risk – providing reasoning alongside every alert. Analysts are never left guessing why an alert has been generated.

Structuring data for optimized analysis

We structure your communications data on ingestion, optimizing it for AI analysis. Not only does this enable accurate AI decision making, but it also helps us manage compute costs, ensuring we can scale to your complex and growing surveillance needs. Structured, single-format data empowers highly performant search capabilities, making it possible to instantly find the information you need when investigating a flagged conversation.

Assured data completeness

Capture and monitor all data types from all your communications channels using Global Relay’s extensive library of data connectors. With over 25 years of experience in communications compliance, our robust reconciliation controls and constant data integrity checks ensure all your data is securely preserved, unalterable, and protected against ransomware. Providing backup assurance, our WORM (write once, read many) storage prevents data modification or deletion, whether accidental or intentional, allowing you to confidently prove compliance with data retention rules.

Global customer support 24/7

With offices in London, New York, and Vancouver, our global support team provides 24/7/365 support whenever you need it.