Eventus, the trade surveillance and financial risk solutions provider, today launched Frank AI, a new artificial intelligence platform designed to transform how compliance teams interact with surveillance data. The new tool embeds generative AI into Eventus’s Validus platform, allowing even non-technical users to conduct complex data investigations using conversational English.
The launch addresses a persistent challenge in financial surveillance: the bottleneck created by cumbersome data queries. Traditionally, when an alert signals potential misconduct, compliance analysts must work with technical teams to build custom queries for deeper investigation—a process that can be slow and iterative.Martina Rejsjö, Eventus Vice President of Product Management, explains, “A major challenge in trade surveillance has been that alerts show only limited data points, forcing users into a slow, iterative process with technical teams to build custom queries for deeper investigation. Frank AI resolves this with a user interface that allows non-technical users to query their data directly. Using a simple chat interface, users can ask questions in plain English. This is possible because Frank utilizes generative AI models specifically trained on the structure of Validus’ data format.”
Frank AI is compatible with leading large language models (LLMs) such as those from OpenAI, Anthropic, and Google. This allows it to leverage the latest advancements in AI to understand user queries and generate the appropriate code to retrieve the requested data.
For example, an analyst can now simply ask: “Show me all cross-market wash trading patterns involving equity and futures for Client XYZ in the past 30 days, including related party analysis.” Frank AI processes the request, analyses data across multiple asset classes, and delivers comprehensive results with a full audit trail.
Security and Data Privacy by Design
A critical concern for financial institutions adopting AI is the security of sensitive client and trade data. Eventus has designed Frank AI to ensure that confidential information remains within the client’s secure environment.
“To ensure data privacy, the client’s confidential data never leaves their hosted environment,” Rejsjö states. “Instead, only the user’s natural language query is sent to the AI model, which returns a piece of code. This code is then executed locally on the client’s own data to generate the required results.”
This approach also means that Frank AI operates within the existing data permissions and governance structures of the firm. “Frank is deployed as part of Validus, so the data used is already within the client’s surveillance system,” Rejsjö adds. “Our surveillance system’s existing permissions guard the data on a per-client basis. Clients have always had access to this data; Frank just gives them a tool to query it more effectively and gain more intelligence.”Deterministic Results for Regulatory Confidence
While generative AI offers powerful new capabilities, its probabilistic nature can be a liability in the highly regulated financial sector, where auditability and repeatability are paramount. Unlike generic AI models that can produce different answers to the same question, Frank AI is designed to be deterministic.
“One issue with generative AI is that it can often give a different answer even when asked the same question, as it interprets your intent differently each time,” Rejsjö says. “Frank AI, on the other hand, produces repeatable queries, providing compliance teams with consistent traceable results. Once you have asked the question, you receive the code. You can always recreate the query with exactly the same results. That’s very important for surveillance because users need to be able to recreate their findings and show an audit trail of how they got an answer. That ability to recreate results is a very significant part of the system as it enables our clients to gain the efficiencies of AI while maintaining regulatory compliance.”
This deterministic approach is crucial for satisfying regulatory inquiries and internal audits, ensuring that compliance teams can trust and verify the results generated by the AI.
By integrating this secure and auditable AI into its powerful Validus platform, Eventus aims to enhance operational efficiency, reduce alert noise, and empower firms to identify nuanced patterns of misconduct more effectively across a wide range of asset classes, including equities, options, futures, foreign exchange, fixed income, and digital assets.
Subscribe to our newsletter
