.jpg)
Financial institutions operate within some of the most demanding regulatory frameworks in the world. Over my career in banking, capital markets and asset management, I have seen that innovation succeeds only when matched by robust controls, clear accountability, and rigorous supervisory engagement.
AI is now being embedded directly into regulated workflows - from compliance monitoring and regulatory interpretation to risk assessment and investment decision-making. The challenge is whether institutions can clearly document how those insights influence regulated outcomes.
This is ultimately a question of compliance architecture:
- How are AI-generated outputs documented and reviewed?
- How is regulatory interpretation validated and challenged?
- How are changes in model behaviour monitored and escalated?
- How can firms show - under supervisory scrutiny - that AI-supported decisions remain controlled, traceable and accountable?
Addressing these questions requires structured AI compliance capabilities.
Institutions seeking to leverage AI in compliance must establish repeatable processes for horizon scanning, regulatory change mapping, gap analysis and documentation of AI-enabled workflows. Compliance teams must be equipped not only to understand AI outputs, but to interrogate them and defend them where required.
I am therefore pleased to join Zango’s Advisory Board at a time when these questions are becoming increasingly urgent. Zango’s regulation-aware agents are designed specifically for regulated financial institutions, supporting horizon scanning, regulatory gap analysis and auditable AI workflows that integrate with existing control structures.
Having seen first-hand at BBVA and other international institutions how supervisory expectations translate into day-to-day operating models, I look forward to helping ensure that AI compliance is grounded in operational reality.
We’re pleased to see this appointment also recognised externally by the Portuguese media, including coverage in Jornal Económico (here).
.jpg)
