The UK’s Financial Conduct Authority is pushing artificial intelligence testing in financial services out of the lab and into controlled real-world environments, with the launch of its AI Live Testing initiative aimed at firms ready to deploy AI systems in production.
The programme is designed for banks, insurers and fintechs that have moved beyond early experimentation and are grappling with how to validate AI safely under live operating conditions.
For QA and software testing teams, the initiative signals a shift in regulatory thinking: AI assurance is no longer about static model validation, but about testing how systems behave once embedded inside real processes, controls and decision flows.
“We’re providing a structured but flexible space where firms can test AI-driven services in real-world conditions, all with our regulatory support and oversight and help from our technical partner, Advai,” explained Ed Towers, head of department in the FCA’s advanced analytics and data science unit.
“Collaboration and communication is at the heart of what we are doing,” he added.
A central motivation behind AI Live Testing is the FCA’s concern that too many AI initiatives stall before production, leaving firms with technically promising systems but no regulatory confidence to deploy them.
“Through live testing we want to help UK innovators move safely beyond ‘POC paralysis’, or what is often described as ‘perpetual pilots’,” Towers said.
“We focus on both quantitative and qualitative factors to get a truly holistic understanding of the AI system.”
– Ed Towers
That concern closely mirrors issues highlighted in QA Financial’s reporting on regulatory QA standards, where supervisors increasingly expect firms to demonstrate evidence of real-world testing rather than relying on theoretical assurances or internal model reviews.
In cross-border work with the Monetary Authority of Singapore, the FCA has already signalled its intent to align AI testing approaches internationally, reinforcing the idea that AI assurance will need to stand up across jurisdictions, not just within internal governance frameworks.
For testing teams, this marks a shift from one-off validation exercises toward continuous, supervised testing that produces artefacts regulators can interrogate.
Testing the whole AI system, not just the model
A key feature of the FCA’s approach is its insistence on defining AI as a system, not a standalone algorithm, a distinction with major implications for QA practice.
“We broadly define the AI system as: the actual AI model, information on the deployment context and core risks … governance and human in-the-loop considerations, evaluation techniques as well as the input and output controls,” Towers said.
That definition pulls together areas that are often fragmented across development, risk, compliance and testing teams.
It also aligns with QA Financial’s coverage of the FCA’s synthetic data governance work, which has stressed that data realism, privacy risk, model behaviour under edge cases and post-deployment monitoring all fall within scope when regulators assess whether AI is being used responsibly.
AI Live Testing follows a structured, three-phase process, discovery, framework validation and AI system testing, combining technical evaluation with supervisory insight.
“We focus on both quantitative and qualitative factors to get a truly holistic understanding of the AI system,” Towers said.
A wider regulatory shift
The FCA’s move lands alongside broader regulatory pressure on testing and resilience across UK financial services.
QA Financial has reported extensively on the Bank of England’s evolving expectations around threat-led testing, cyber recovery and operational resilience, where firms are increasingly required to demonstrate that systems work under stress, not just on paper.
Taken together, these initiatives point to a convergence in supervisory thinking: whether testing AI models, digital platforms or recovery capabilities, regulators are prioritising observable behaviour, repeatable evidence and live-environment validation.
For QA and software testing teams inside banks and insurers, AI Live Testing is less about experimentation and more about accountability.
It reinforces a future where AI assurance sits alongside resilience testing, security testing and operational scenario testing, with QA teams expected to provide the evidence that AI systems are safe, governed and ready for real customers, not just internal demos.
COMING IN 2026
Why not become a QA Financial subscriber?
It’s entirely FREE
* Receive our weekly newsletter every Wednesday * Get priority invitations to our Forum events *
REGULATION & COMPLIANCE
Looking for more news on regulations and compliance requirements driving developments in software quality engineering at financial firms? Visit our dedicated Regulation & Compliance page here.
READ MORE
- AI adoption strains JPMorgan testing
- Can banks ‘outsource’ AI accountability?
- HDFC Bank raises testing stakes
- Is observability banking QA’s next discipline?
- Barclays on AI testing, telemetry and kill switches
WATCH NOW
QA FINANCIAL PODCASTS
