TL;DR
The financial services sector has shifted from theoretical exploration to regulated, real-world execution, highlighted by Europe's first live autonomous payment and the establishment of comprehensive regulatory blueprints. While tier-one institutions co-develop sovereign infrastructure to maintain strict localized compliance, the industry is adopting automated, real-time assurance platforms to mitigate the risks of autonomous decision-making.
Systematic Regulation and the Rise of Automated Assurance
Financial regulators and compliance technology providers are shifting from static, post-facto compliance reviews to real-time, automated guardrails to govern autonomous financial systems.
"Our work in the Supercharged Sandbox with Agent Assure creates a new approach, delivered live and using small language models to act as independent assessors of primary models. It has been designed around the principle of unified assurance." — Aveni's Agent Assure Pilot in Regulatory Frameworks and Liability for Agentic Finance
As financial institutions transition to highly autonomous operational modes, traditional point-in-time validation fails to account for live drifts and personalized customer journeys. Deploying secondary, domain-specific systems like FinLLM to continuously inspect primary transaction networks represents a critical path to proving "meaningful human control" and avoiding personal liability under the UK's Senior Managers Regime Regulatory Frameworks and Liability for Agentic Finance. This shift is further reinforced by the FCA's 147-page Mills Review, which establishes a five-level autonomy spectrum to govern retail financial services through 2030 Regulatory Frameworks and Liability for Agentic Finance
.
What to watch: How the Financial Conduct Authority implements its recommended supervisory framework to move from episodic, document-based oversight to continuous, systemwide risk monitoring Regulatory Frameworks and Liability for Agentic Finance.
Sovereign Infrastructure and Autonomous Transaction Execution
Tier-one global banks are bypassing standard commercial software vendors to co-develop sovereign AI infrastructure while simultaneously proving the feasibility of direct machine-to-machine transactions.
"Together with Santander, Presight intends to build an intelligence platform designed to bring that experience within reach of every customer: an AI-powered financial agent that knows them, anticipates them and acts for them within the boundaries they set." — Santander and G42 MoU Announcement in Banco Santander's Sovereign AI Strategy
By combining sovereign infrastructure from G42 with regulated payment rails like Mastercard Agent Pay, legacy banks are establishing a secure foundation where non-human actors can execute real-world financial transactions for 176 million global banking customers Banco Santander's Sovereign AI Strategy. This operational push is backed by substantial capital, with Santander deploying a $400 million venture fund through Mouro Capital to control both the underlying technology and the early-stage startup ecosystem that feeds into it Banco Santander's Sovereign AI Strategy
.
What to watch: The deployment of localized, self-hosted financial systems across Santander's retail network in Europe and Latin America through G42's Catalyst platform Banco Santander's Sovereign AI Strategy.
What surprised us
- Europe's first live non-human payment was executed within a regulated framework: In March 2026, Santander and Mastercard completed a live end-to-end payment using Mastercard Agent Pay, proving that autonomous non-human actors can securely initiate and complete transactions within predefined account limits Banco Santander's Sovereign AI Strategy
.
- The extreme gap between autonomous ambition and compliance readiness: While 99% of financial firms plan to deploy autonomous systems, a mere 2% report having adequate guardrails in place, highlighting a massive, underserved market for automated assurance software Regulatory Frameworks and Liability for Agentic Finance
.
- Regulators are preparing to regulate themselves with AI: The FCA's Mills Review explicitly recommends that the regulator build its own supervisory framework, shifting its own operations from episodic, document-based reviews to continuous, systemwide risk oversight Regulatory Frameworks and Liability for Agentic Finance