As financial markets become more fragmented, faster-moving, and data-saturated, legacy compliance systems are buckling under the pressure. Solidus Labs Co-Founder and CXO Chen Arad shares how a new paradigm—Agentic AI—is giving compliance teams the firepower they need to regain control, cut through complexity, and finally get ahead of financial crime.
Agentic AI refers to a type of artificial intelligence that focuses on creating autonomous systems capable of making decisions and performing tasks without constant human intervention.
Chen Arad is the Co-Founder and Chief External Affairs Officer at Solidus Labs, where he leads global market engagement, policy strategy, and regulatory partnerships. He serves on the CFTC’s Global Markets Advisory Committee and is a sought-after speaker on market integrity, financial crime prevention, and digital asset regulation.
Digital Journal: Why are so many compliance systems still failing despite record-high investments in technology and talent?
Chen Arad: It’s not a failure of effort, it’s a failure of architecture. Despite the billions poured into technology and hiring, compliance teams remain buried under alert fatigue, disconnected systems, and a shortage of experienced analysts. Ultimately it’s because legacy risk monitoring systems were built to support a 1990s-style market—centralized, slower-moving, institution-driven, and focused on a narrow set of asset classes.
Today’s reality is vastly different. Markets are hyper-fragmented and operate 24/7. Retail participation and off-exchange trading has exploded, carrying heightened risks cyber-enhanced financial crimes and cross-product and cross-market manipulation. Additionally, firms are increasingly exposed to digital assets—where risks move faster, signals are harder to interpret, and abuse often spans multiple platforms and jurisdictions. And yet – most compliance systems haven’t kept up.
We’ve seen firsthand how tech sprawl and siloed data lead to blind spots that bad actors exploit—costing firms not only thousands of hours but, in some cases, billions in fines and losses. And even worse, it doesn’t solve financial crime effectively – To give a few exmaples – only 3% of US dollars laundered ever get identified; Only 4% of SARs lead to enforcement meaning more than 90% of the reports by compliance teams aren’t about real crime; 20% of M&As and 5% of quarterly earning reports in US stock markets are accompanied by insider trading
DJ: What is Agentic-Based Compliance, and how does it work?
Arad: Agentic-Based Compliance is a complete reimagining of how compliance operations function. Rather than relying on static rulebooks and pre-scripted workflows, we’ve introduced a dynamic system built around a fleet of autonomous, intelligent AI agents, supervised by human compliance professionals.
Each agent is goal-oriented and specialized—it might detect spoofing, enrich alerts, gather OSINT, or triage cases by severity. But what makes the system truly transformational is the way these agents collaborate—analyzing real-time data, learning from patterns, and continuously coordinating to produce faster, smarter, and more precise outcomes. It’s not just automation. It’s autonomy at scale—with full human oversight.
DJ: What makes Agentic AI a necessary leap forward from traditional surveillance software?
Arad: Most surveillance software on the market today is still rules-based and rely on hardcoded, static logic. It automates tasks but waits to be prompted. That model might anecdotally reduce some friction, but it doesn’t change the game—it still demands a huge lift from compliance teams.
Agentic AI takes a fundamentally different approach. These agents proactively scan for risk, adapt to new behaviors, and autonomously investigate suspicious activity. Think of them not as tools, but as virtual teammates—AI analysts that think, plan, and act at machine speed, without ever compromising human control or regulatory oversight. It’s a 100x upgrade in efficiency, accuracy, and scalability.
DJ: How can an innovation born in crypto be applied to traditional finance markets?
Arad: Crypto gave us a proving ground like no other—hyper-fragmented, retail-first, 24/7, and ripe with cyber-enhanced manipulation. We built Agentic AI to handle that chaos—combining Wall Street rigor, crypto-native innovation, and cybersecurity dynamics into high-velocity detection technology. If you can make surveillance work in crypto, you can make it work anywhere.
Now, as the lines between TradFi and digital assets blur, the challenges are converging. TradFi is facing the same multi-venue fragmentation, data volume, and threat complexity. Agentic-Based Compliance isn’t crypto-native or TradFi-native—it’s market-native, built for today’s hybrid reality.
DJ: Can Agentic AI prevent complex manipulation events like the recent $700M penny-stock fraud in Japan?
Arad: Yes—and that’s exactly why we built it. These schemes are rarely isolated. They unfold gradually, across markets and asset classes, driven by subtle shifts in behavior that legacy systems consistently miss. Outdated models might detect the symptoms. Agentic AI uncovers the root cause.
Our agents are designed to surface anomalies across accounts, track evolving behavior patterns, correlate external signals—from news headlines to social media—and even trace coordinated actions across trading desks or entities. It doesn’t just flag problems; it assembles the full narrative, giving investigators a real head start in stopping manipulation before it spreads.
DJ: What role do human analysts play in this new model?
Arad: A central one. This isn’t about replacing human analysts—it’s about unlocking their full potential and letting them focus on what they signed up for – protecting investors and fighting crime. Agentic-Based Compliance is designed to relieve teams of the manual, repetitive, and disconnected tasks that burn them out and slow them down.
Instead of spending hours stitching together fragmented data or drafting investigation case reports from scratch, analysts are equipped with pre-enriched, fully contextualized insights—delivered by their AI teammates. This allows them to focus on what truly matters: conducting deeper investigations, making informed judgment calls, and driving strategic decisions with the confidence that nothing critical has been overlooked.
