I’ve joined Tangos as Chief Evangelist, and I did it for one reason.
In 24 years in this industry, I have struggled with the lack of controls that work effectively to identify crime as it evolves in real time. As far back as 2014 I had serious concerns about being required to adopt rules- and scenario-based systems because regulation demanded them, when in fact they weren’t fit for purpose. A list of controls based on yesterday’s understanding of risk isn’t a control. It’s a gateway for illicit actors to use new, undetected mechanisms. Covid opened our eyes to cyber security risk, and criminals quickly realised that the fragmentation of data and connections across financial institutions was ripe for exploitation.
We need AI in financial crime risk management more than we realise. We have no choice. But the truth is that the vendor market is very fragmented. Many providers talk about AI. Fewer have fully understood what AI can do. Fewer still have actually started to use it in their products beyond roadmaps and proofs of concept, often because they don’t have the data to test it even in beta. And then there is the last group: those that have AI genuinely working in risk management.
That last group is small, and Tangos is the first time I have seen pure AI, not just an LLM, actually working. That excites me. Tangos is currently a financial crime investigation platform, focused on the work that starts after an alert and built to take an investigation through to resolution, with the evidence, reasoning and case file needed to support the final decision. But that is only the tip of the iceberg. The roadmap we are building will create a new category of risk management controls: AI-native, and continuously updated as new and emerging typologies are fed in from a wide range of data sources.
So the answer to why I joined is simple. This is the future of financial crime risk management, and I want to be part of it, bringing my experience to make a real difference and leave a legacy that improves how risk is managed.
My background in financial crime compliance and risk management
I started out in law, and a legal traineeship at the European Central Bank taught me early that regulation reads very differently on paper than it does inside an institution. If you don’t understand risk, you can’t apply regulation. It’s that simple. Go back to the fundamentals: products, services, clients, third-party exposure, geography. Those risk assessments, and the knowledge behind them, are fragmented across organisations. Yet it is where those risks come together that financial crime actually originates.
Needing to learn those fundamentals from the ground up is what took me into banking. I spent years inside global financial institutions, from London to Singapore to Hong Kong: product development at JP Morgan, public- and private-side advisory and policy at BTMU (now MUFG), building the global Control Room at ING and at Standard Chartered, then private banking, correspondent banking and trade finance. I went on to lead HSBC’s financial crime remediation out of Hong Kong for Asia Pacific, and finally served as Citi’s APAC Head of Regulatory and Financial Crime Compliance Internal Audit. Literally across all three lines of defence.
Building GRC frameworks and running remediation are both humbling work. You see exactly how controls that look effective on paper fail in practice, and how often it is ineffective, inefficient process that lets financial crime through. Nobody sets out to fail. They are drowning in manual work. You also learn what works. It is a never-ending process of assessing effectiveness and prioritising risk, because you can’t cover everything. You have to accept some risks because the probability and impact are acceptable. But with the big ones, a single failure brings legal, regulatory, financial and consumer consequences, and reputation and trust with the public and regulators are very hard to win back.
I founded Virtual Risk Solutions in 2015 to bring that practitioner view to firms across banking, fintech and digital assets. It only really took off in 2020, when Covid hit. I had built a completely virtual financial crime consultancy and been told it would never work. Covid proved that it did. That same year I founded RAW Compliance, a social enterprise, to give practitioners free training that is practical rather than theoretical. More than 17,000 professionals have trained with us since.
Along the way I have published nine books, with two more coming in the next few weeks (and that’s it for the foreseeable future!), launched and hosted The Compliance Word podcast, and spoken on more conference stages than I can count. My relationship with this industry hasn’t changed. I genuinely love it, I believe we can make a difference, and I don’t let anyone off the hook when I think we can do better, including myself.
Why I chose Tangos for AI-powered financial crime investigations
I approached Tangos the same way I would approach any provider in this market. I wanted to understand what it actually did, how it would work inside a complex financial institution, and whether it would genuinely improve an investigation.
I am very wary of technology promises. I’ve sat through the demos. I’ve watched tools get bought and then quietly worked around, or used for only part of their functionality with other products bolted on. Those are all recipes for disaster. A polished demo and a useful compliance tool are not the same thing. A system has to work in the reality of financial crime operations, where decisions need to be trusted, explained and defended.
What convinced me was that Tangos understands the problem, and has the right people to build AI that actually solves it. It is a team that has done this work themselves, and that makes the difference. How the product looks isn’t the first consideration (sorry, Steve Jobs). What matters is whether it can defend a decision. And I have always believed that what people really invest in isn’t only the technology or the idea. Those matter, but people also want to work with a company that is as invested in their success as they are. That is what brought me to Tangos.
How Tangos closes the financial crime investigation gap
Its current focus is a good example. Financial institutions have invested heavily in transaction monitoring, screening, analytics and detection, and alert generation has improved. We lived for years with detection accuracy of 2-8%, and better calibration is slowly moving that on. (I’m still not convinced about the data feeding these systems, but that’s a separate discussion.) Yet behind almost every meaningful alert sits a deeply manual, time-consuming investigation.
Someone still has to establish what happened. Someone has to test competing explanations, find and validate evidence, understand relationships and reach a conclusion that can withstand internal and regulatory scrutiny. And every person will assess it through a different lens, risk view and perspective. That isn’t wrong. It’s human, and training doesn’t remove it. It’s bias.
That investigation gap is where Tangos is focused. Tangos takes the alert and performs the investigation. It develops hypotheses, gathers and assesses evidence across sources, jurisdictions and languages, and produces source-traced, auditable case files. These go to the analyst, who assesses the recommended next steps against the institution’s risk appetite. An investigation is not a pile of search results. It is the work of deciding what the evidence means, weighing different explanations and documenting why a conclusion was reached.
Why does this matter now? Because most financial institutions do not need to replace the systems, processes and expertise they already have. They need a better way to turn an alert into an investigation they can stand behind: a defensible decision, where the control and the decision-making process can be fully reproduced for a regulator or law enforcement.
Tangos works alongside existing financial crime infrastructure. It is not there to remove experienced practitioners from the decision. The final call stays with people who understand the institution, the risk and the consequences. The aim is not an over-engineered answer that looks impressive on paper. It is an outcome that can be trusted, reasoning that can be followed, evidence that can be traced and, most importantly, a decision that can be explained clearly to someone who was not in the room when it was made.
What I’ll be doing at Tangos: financial crime AI, governance and implementation
As Chief Evangelist, I will lead Tangos’ engagement with the financial crime community and help translate developments in AI and investigation technology into practical tools for financial institutions, regulators and practitioners.
My job is to keep the practitioner’s voice in the room: to make sure we really understand the problem, to create AI use cases that make a credible difference in financial crime risk management, and to be honest with the financial crime and wider risk management community about what works and what doesn’t, what is possible today and what isn’t yet. I don’t think anything is impossible. Some things just haven’t been done yet, and that is a very different proposition.
There is a lot of noise around AI in financial services, so I want these conversations to stay grounded in governance, explainability and the realities of implementation. What does good governance look like? What should investigators be able to see and challenge? What does an institution need in order to explain and defend the outcome? Those are practical questions, not theoretical ones.
I have already authored Tangos’ forthcoming Financial Crime Trends and Threats 2026 report and will lead its 2027 predictions publication. Both follow the same approach: what is changing, what it means in practice, and where the financial crime community should be asking harder questions.
Eyal Azoulay, Tangos’ CEO, put it this way:
“Oonagh has spent her career in the rooms where financial-crime decisions get made – alongside investigators, compliance leaders and regulators. She knows what those teams need because she’s sat in their seat, and she’s invested years in how AI changes compliance work. That’s the caliber Tangos is built on: every function of this company is staffed with top talent and proven success in its field.”
Why the next phase of financial crime AI is about proof
A brilliant demonstration means very little if an investigator cannot trust the outcome, explain the decision or defend it when a regulator comes asking. The next phase of this market will be about proof.
For me, that means looking beyond whether AI can do something and asking whether it improves the investigation that follows an alert. Whether the evidence is stronger. Whether the reasoning is clear. Whether the case file stands up to scrutiny. Whether experienced practitioners stay in control of the decision.
That is the standard I intend to bring to this role. I’ll be sharing more of this work through Tangos and in my conversations with the financial crime community. If these are the questions you are working through too, follow along or connect with Tangos.