Sonja Davidovic

Senior Digital Expert on Transformative Technology

International Monetary Fund

View the agenda

Sonja Davidovic is a Senior Expert in Digital Assets at the International Monetary Fund (IMF), working at the intersection of digital finance, cross-border payments, and agentic AI. Previously, she led the digital asset program at the BIS Innovation Hub Singapore Centre, spearheading pioneering initiatives including Projects Mariana, Mandala, and Rialto. With experience spanning the IMF, BIS, Monetary Authority of Singapore, and European Central Bank, she combines deep policy expertise with hands-on innovation to help shape the future of global finance.

Speaking At

9:30–11:00amMon 16 NovLaunchpadLaunchpad RoomOpen
Others

Trust Needed: The Future of Agentic Payments - Powered by Amazon

LaunchpadLaunchpad RoomOpen
About this session

_When an AI agent makes a payment across borders, who sets the rules and who is accountable if something goes wrong?_
This session introduces the Agentic Payments Trust Principles, developed through the FCA Supercharged Sandbox as well as perspectives across the payment value chain will explore what it takes to trust AI agents with our payments – and how the trust principles can work in practice.
A panel discussion with GFIN AI Compass members and other regulators will examine how the trust principles connect with emerging approaches to agentic payments around the world, and what cooperation is needed to make this work with confidence.
Attendees will gain insights into the trust principles in-progress, how they have been applied , and the cooperation needed to govern AI agents in cross-border transactions.
9:30am - 9:45am
Opening Keynote
Speaker(s): Simran Singh, AI Lab Lead, Financial Conduct Authority (FCA), Amira Karim, Head of International Public Policy, Payments & Financial Services, Amazon, Nicole Sandler, Chief Ecosystem Officer, Ubyx Inc
.
9:45am - 10:20am
Stakeholders’ Perspectives: Connecting to the Ground
Speaker(s): Jiang Fangfang, Asia Lead for Digital Financial Services, International Finance Corporation, Shayan Hazir, Chief Digital Officer, Asia (xHK) & MENAT, HSBC, Vishal Mehta, Managing Partner, Travel & Payments, Nicole Sandler, Chief Ecosystem Officer, Ubyx Inc (Moderator)
.
10:20am - 10:55am
Panel Discussion: How the Trust Principles Link to What Various Jurisdictions are Thinking About in Relation to Agentic Payments
Speaker(s): Jose Manuel Marques Sevillano, Director Financial Market Infrastructures & Innovation, Banco de España, Sonja Davidovic, Senior Digital Expert on Transformative Technology, International Monetary Fund, Simran Singh, AI Lab Lead, Financial Conduct Authority (FCA) (Moderator)
.
10:55am - 11am
Audience Q&A and Closing
Speaker: Simran Singh, AI Lab Lead, Financial Conduct Authority (FCA)
2:00–3:30pmMon 16 NovPublic-Private RoundtableRoundtable Room 3Open
The Scaling Challenge in Tokenized Markets

GL1 Dialogue: Infrastructure for a Programmable Financial System

Public-Private RoundtableRoundtable Room 3Open
About this session

Global Layer One (GL1) has brought together policymakers, financial institutions and infrastructure providers to advance an ecosystem of open, interoperable and trusted digital infrastructures. This collaboration has produced practical outputs such as the Programmable Compliance Toolkit, which provides a common approach to implementing compliance across tokenized asset systems, and the Market Infrastructure Toolkit, which is already being used by market infrastructure providers to assess and demonstrate alignment with internationally recognised standards and risk management principles.
This roundtable convenes the GL1 community to share experiences from adopting and applying these toolkits, including practical implementation lessons and considerations for broader adoption. Participants will also look ahead to how this work can be extended to support new use cases and capabilities as the industry progresses towards a programmable financial system.
9:45–11:15amTue 17 NovPublic-Private RoundtableRoundtable Room 4Open
Accountability & Liability In Agentic Systems

Multi-Agent AI in Financial Systems and the Implications for Central Banking

Public-Private RoundtableRoundtable Room 4Open
About this session

Advances in AI are enabling multiple intelligent agents to collaborate, make decisions and execute complex tasks with increasing autonomy. In financial systems, these capabilities could support new approaches to economic simulation, policy analysis, supervisory monitoring and increasingly autonomous operations.
While offering significant opportunities to enhance decision-making and system efficiency, they also introduce new questions around governance, validation, accountability, systemic behaviour and financial stability. This roundtable will critically assess these emerging capabilities, associate risks and examine the implications for central banking.
1:30–3:00pmTue 17 NovPublic-Private RoundtableRoundtable Room 2Open
The Scaling Challenge in Tokenized MarketsGlobal Frameworks and Shared Best Practices in Payments Chains

Programmable Compliance: Which Rules Travel with the Money, and How?

Public-Private RoundtableRoundtable Room 2Open
About this session

As money becomes programmable, across central bank digital currencies, tokenized deposits, and stablecoins, a real question is emerging about where compliance should live: embedded directly in the asset as a rule that travels with it, layered on top through richer messaging and real-time screening, or enforced separately by whichever party touches the transaction last.
Players are all building credible but different answers in parallel, and none has converged on how these should interoperate across jurisdictions with different compliance regimes. The stakes are rising on every front — real-time cross-border settlement, tokenized securities, and increasingly, autonomous agents initiating transactions on their own — all of which need a compliance check at the moment value moves, not after. What would each of these approaches to programmable compliance actually need to interoperate, and which rules win when they conflict across borders?
Discussion questions:
1\.         Should compliance logic live in the asset itself, in the messaging/settlement layer, or in a separate real-time screening service, and does the answer change depending on whether it's a CBDC, a tokenized deposit, or a stablecoin?
2\.         When a transaction crosses borders, whose compliance rule actually governs — sender's jurisdiction, receiver's, or the rail's — and what happens when they conflict?
3\.         Retrofitting compliance onto existing settlement infrastructure and building it in from scratch are both live strategies right now — what would each need to prove to become the default, and is convergence between them actually possible?
4\.         As more transactions — including those initiated by autonomous agents — require a compliance check the moment value moves, what are the minimum interoperability standards needed for rules to actually travel with money across rails?

Speakers

Sonja Davidovic

Speaking At

Trust Needed: The Future of Agentic Payments - Powered by Amazon

GL1 Dialogue: Infrastructure for a Programmable Financial System

Multi-Agent AI in Financial Systems and the Implications for Central Banking

Programmable Compliance: Which Rules Travel with the Money, and How?