At XLoD Global London, the most revealing conversations did not start with product demos or panel soundbites. They started with harder questions.
Can we trust the data behind our alerts? Can we explain why AI made a decision? Can we prove that surveillance is complete across every channel? Can we show regulators not just that controls exist, but that they work?
Those questions defined the mood in London.
Across sessions, booth conversations, presentations and the informal discussions that often happen between the official agenda, one thing became clear: the compliance conversation has moved on. The industry is no longer simply asking how to automate more, review faster or reduce false positives. Those challenges still matter, of course. But the bigger issue now is confidence.
In a market shaped by rising regulatory pressure, growing operational complexity, fragmented data and accelerating AI adoption, firms are searching for something more durable than efficiency. They are looking for regulatory certainty.
That was the thread running through NICE Actimize’s experience at XLoD Global London 2026, where we joined risk, compliance, technology and regulatory leaders to discuss the future of non-financial risk. The event confirmed what we are hearing across the industry: compliance leaders are ready to modernize, but only if modernization is explainable, defensible and built on trusted data.
AI Has Moved From Curiosity to Consequence
Unsurprisingly, AI dominated much of the conversation. But the tone was different from the AI discussions of the past few years.
The industry has moved beyond broad excitement and speculative use cases. The questions now are more practical, more urgent and more mature. Where can AI create real value today? Where does it need stronger governance? Where can it safely assist human decision-making? And where are firms still right to be cautious?
As Cromwell Fraser, Director - SME, NICE Actimize, observed, “AI is the main topic of conversation,” especially when it comes to governance, data completeness, data lineage, false positive management, agentic support for compliance teams and AI-driven detection.
That range of topics says a lot about where the market is heading. Compliance teams see clear potential for AI to reduce noise, improve prioritization, support alert triage and help analysts work faster and smarter. But when the conversation turns to AI-led detection in high-stakes surveillance environments, the mood becomes more measured.
The appetite is there. The scrutiny is, too.
Firms want AI that can identify behaviors and patterns that humans may miss. But they also want to know whether those outcomes are repeatable, explainable and robust enough to stand up to internal challenge and regulatory review. In compliance, “interesting” is not enough. The output has to be trusted.
That same pragmatism came through in conversations with compliance leaders on the show floor. James Chidwick, Account Executive, FMCC, Sales Channel Management, NICE Actimize, noted that many of the directors of compliance and surveillance officers he spoke with were focused less on abstract AI ambition and more on immediate operational pressure: ensuring data completeness, using AI to reduce alert noise so analysts are not overwhelmed and improving the reliability of transcription across multiple languages.
Those priorities point to where AI can have its most immediate impact. For firms dealing with high alert volumes, fragmented communications channels and increasingly global operations, AI is not just a future-state concept. It is becoming part of the practical toolkit for helping teams focus attention where it matters most — provided the data is complete, the outputs are explainable and the human reviewer remains firmly in control.
“Because the AI Said So” Is Not a Compliance Strategy
That tension was central to NICE Actimize General Manager Konstantinos Rizakos’s XLoD presentation, “Achieving Regulatory Certainty in an AI World.”
His message was simple, and it landed because it reflects the reality compliance teams are now facing: “Because the AI said so” will not satisfy users, risk teams, auditors, senior management or regulators.
Not every AI use case requires the same level of explanation. AI used for research, process design, summarization or workflow support may not always require deep reasoning transparency. But once AI moves into production, especially in areas such as surveillance, conduct risk, market abuse detection or regulatory decision support, the bar raises.
At that point, firms need to show why a system reached a decision, not just what the decision was.
That distinction matters. Traditional explainable AI techniques have helped the industry make progress, but they are not always sufficient. In some cases, they may be too limited, too model-centric or not reliable enough to satisfy the demands of production compliance environments.
A more advanced capability is now emerging: counterfactual explanations. These explanations help answer a more precise and useful question: what would have needed to change for the AI system to reach a different conclusion?
For compliance teams, that is a major step forward. Counterfactual explanations can help move AI outputs from opaque recommendations to outcomes that are more mathematically robust, more defensible and more actionable. They give firms a clearer way to understand the evidence behind AI-driven decisions and a stronger foundation for regulatory engagement.
In other words, the future of AI in compliance will not be defined by who deploys it first. It will be defined by who can explain it best.
The Near-Term Opportunity: AI as the Investigator’s Advantage
While the industry is still cautious about full AI-driven detection, there is growing confidence in AI as an investigative assistant, especially where it can reduce alert noise, improve prioritization and help analysts avoid being overwhelmed by volume.
Paul Cottee, Director, Regulatory Compliance, NICE Actimize, captured the mood well: “The feeling is definitely ‘where can we employ these tools?’ but not that it is the magic bullet.”
That is exactly the right framing. AI is not being viewed as a replacement for compliance professionals. It is being viewed as a way to sharpen their judgment, speed up their work and expand the context available to them.
One example discussed at the event was the use of AI to support investigations into trading behavior. If a trader triggers an alert for unusually heavy activity at a certain time of day, AI could help review that trader’s historical activity and determine whether the behavior is genuinely unusual or part of a broader pattern. That does not mean AI is making the final call. It means AI is helping the investigator ask better questions faster.
That distinction is important because it reflects where many firms can create value now. AI can help reduce false positives, prioritize alerts, identify relevant context, summarize investigative material and surface patterns across large volumes of data. It can make compliance teams more efficient, but also more effective.
The real opportunity is not automation for automation’s sake. It is better human decision-making, supported by better intelligence.
That was one of the clearest messages from XLoD London: AI is no longer just an innovation topic for compliance teams. It is becoming an accountability topic.
But AI is only one part of the regulatory certainty equation. The next challenge is just as important: whether firms can connect surveillance, conduct, culture and data into a compliance model that regulators can trust.
In part two, we’ll look at why surveillance silos are collapsing, why data trust is becoming a regulatory issue and what compliance leaders need to build next.
Missed us at XLoD London? Learn how NICE Actimize helps firms strengthen compliance oversight, reduce noise, support better decision-making and move toward greater regulatory certainty.
Explore our solutions here.