Signal Over Noise: Rethinking Trade Surveillance (The Demo Room #18)

Welcome to The Demo Room – your front-row seat to the future of RegTech, RiskTech, and AI innovation. 

In this series, we document our research interviews with the most forward-thinking vendors tackling the industry's biggest challenges. Each blog is built around a comprehensive product demo, providing clear insights into how these innovations address industry challenges.

On this occasion, we feature Eventus, provider of Validus, a multi-asset trade surveillance platform built for firms operating across high-volume and fast-evolving markets.

---

Market abuse is rarely visible from a single trade. Behaviours such as spoofing, insider dealing, marking the close and cross-product manipulation only become legible when orders, executions, cancellations, price movement, trader behaviour and venue activity are connected in the right way.

That connection depends on data. And data is where most surveillance frameworks start to break down. Poor quality, inconsistent formats, fragmented systems and miscalibrated thresholds mean that surveillance teams are often working with a weakened foundation before an analyst has looked at a single alert.

The challenge is compounded by rising regulatory expectations. Supervisors increasingly want to see that surveillance logic works in practice: that data is complete, thresholds are appropriately calibrated, investigations are consistent and decisions can be defended.

It’s a demanding operating environment. More data, more products, more venues, and continued pressure to reduce noise without weakening the control framework.

The Problem for Firms

Surveillance teams are not short of alerts. They are short of signal.

Legacy systems can often detect known patterns, but they struggle with volume, data complexity and adaptation to new markets. Poor threshold calibration creates blind spots in one direction and excessive alert volumes in the other. The result is a function that is simultaneously over-stretched and under-informed.

Data quality sits underneath the entire control. If orders, fills, cancellations, market data, reference data, account identifiers and venue feeds are not mapped consistently, the surveillance logic is weakened before an analyst has looked at a single alert. In high-volume environments, even small inconsistencies in the data foundation compound quickly.

“There’s only a handful of platforms that can handle serious volume and data complexity.” – Joseph Schifano, Global Head of Regulatory Affairs, Eventus

Fragmentation makes the issue worse. In many firms, trade surveillance, eComms surveillance and trade reporting are still managed by different teams using different tools. That makes it harder to reconstruct events, identify intent and evidence why an alert was closed or escalated.

Regulatory expectations have also moved on. Supervisors are focusing on whether surveillance frameworks work in practice. Firms need to show that data is complete, logic is appropriate, thresholds are governed, investigations are consistent and alert handling can be defended. A surveillance function that generates alerts but cannot explain its decisions is not meeting the bar.

The challenge is not just detecting more. It is making better use of what the data is already telling you.

A Solution: Validus by Eventus

Validus combines configurable detection logic, alert workflow, automation, visual investigation tools and AI-enabled data interrogation. It supports surveillance across traditional and digital asset markets, with procedures covering market manipulation, risk controls and related use cases.

1. Flexible data ingest and normalisation

Validus is built around a practical reality: clients do not all provide data in the same format. Eventus normalises client data from multiple environments and maps it into the structure needed for consistent surveillance.

This is a major part of the product’s value. Trade surveillance depends on the integrity of the data foundation. Firms need confidence that trading activity, market data, order lifecycle events and reference data have been mapped accurately before surveillance logic is applied.

“What I find very compelling with Validus is that the data, the underlying data, is so accessible.” – Martina Rejsjö, Head of Product Strategy, Eventus

That accessibility matters because surveillance teams need to move beyond static alert fields. They need to interrogate underlying activity quickly, ask follow-up questions and understand whether the alert reflects meaningful behaviour.

2. Alert workflow built around investigation

Validus centres on the alert workflow, with tools for status, tagging, escalation, continuity and resolution. Analysts can move from a high-level alert dashboard into more detailed views showing order lifecycle, account behaviour and supporting evidence.

The Market Visualizer shows trading activity against market data, including orders, executions, cancellations and price context. Insider dealing workflows can bring together price movement, news events, transaction timing and potential profit.

This helps analysts move from a triggered alert to the surrounding evidence more quickly. The objective is a better decision, backed by a clearer record.

3. Frank AI and plain-language access to surveillance data

Frank AI is one of Validus’ most distinctive features. It allows users to query Validus data in plain language and receive charts, tables and code-backed outputs.

The value is straightforward. In many surveillance environments, analysts can only go so far before they need technical support to query underlying data. Frank lowers that dependency by turning ordinary questions into executable queries against the client’s own Validus data.

This is particularly useful for follow-up investigation. A user can ask which accounts traded one-sided in the final five minutes of the trading day, request a chart, or narrow results to a specific window or behaviour. Frank is trained around Validus’ data model and trade surveillance concepts, including orders, executions, cancellations, account activity and trading windows.

The governance design is also important. Client data remains within the hosted environment, while outputs are supported by visible code. Users can inspect the query that produced the result and reuse it where repeatability is required.

That matters in surveillance, where a useful answer is not enough. Firms need to understand how the answer was produced and defend the process if challenged.

4. Automation with accountability intact

Validus supports automation across alert workflow, tagging, escalation, continuity and low-value alert reduction. Eventus also supports client-specific automations layered on top of base procedures.

This allows firms to codify repeatable analyst logic. For example, a firm may define conditions under which certain low-relevance alerts can be closed before reaching an analyst, with a standard explanation and audit trail.

This is a pragmatic approach to automation in market conduct. Human oversight remains central, while repetitive work can be reduced where the firm has already approved the logic.

5. Cross-product and emerging-market trade surveillance

Cross-product manipulation is becoming more important as bad actors exploit relationships across instruments, venues and asset classes. Validus is developing cross-product surveillance through normalised data and correlation-based logic, allowing the platform to identify behaviours such as spoofing across related products.

Eventus is also preparing for newer market structures, particularly prediction markets and blockchain-derived data.

Prediction markets create event-based instruments where activity may be driven by news, social sentiment, political developments or highly specific catalysts. Digital assets and tokenisation extend the surveillance perimeter further, requiring firms to think about on-chain and off-chain activity together.

Validus’ ability to ingest, normalise, visualise and interrogate complex datasets is increasingly relevant beyond traditional exchange-traded products.

Parker & Lawrence’s View

One of Validus’ strongest qualities is data accessibility. Surveillance teams need to move quickly from a triggered alert into the underlying evidence. Validus is built to support that, with visual investigation tools, configurable logic and plain-language data interrogation through Frank AI. Critically, every Frank AI output is backed by visible, auditable code. In a regulatory context, how an answer was produced matters as much as the answer itself.

Adoption of AI in surveillance is constrained not by appetite but by governance, validation and internal model risk requirements. Eventus’ approach is well calibrated to that reality. AI is applied where it improves access and reduces technical dependency, without displacing accountability from the surveillance function.

As markets grow more complex and regulatory expectations around defensibility continue to rise, the premium will be on platforms that support both detection and investigation at scale. Validus is well positioned for that.