The definition of research and development (R&D) in financial services has long been narrowly interpreted – focused on software development, trading platforms, and risk engines. However, HMRC’s post-2024 reforms, which now include data, cloud computing, and pure mathematics as qualifying fields, signal a major shift. Financial institutions are being encouraged to view innovation not as isolated software builds, but as interconnected eco-systems where computation, regulation, and human decision-making evolve together.

From code to cognitive R&D in financial services

Traditionally, R&D claims were based on solving engineering problems, making systems work faster or more reliably. Today, the scope has expanded to include cognitive challenges, ie, how systems think, adapt, and learn.

Projects that qualify now include:

• Developing new stochastic models for credit risk using advanced mathematical methods.
• Creating novel numerical solvers for real-time option pricing under complex volatility conditions.
• Designing unsupervised learning models to enhance anomaly detection in transaction monitoring.

These initiatives go beyond software upgrades, they represent efforts to resolve scientific and technological uncertainty, aligning with the Department for Science, Innovation and Technology (DSIT) definition of R&D.

LLMs and systems thinking in financial services R&D

The rise of large language models (LLMs) has accelerated this transformation. Financial firms are now building cognitive infrastructures, not just codebases.

Examples of qualifying R&D include:

• Creating domain-specific LLMs for risk and compliance (eg translating complex FCA directives into executable policy controls).
• Embedding explainability and traceability layers to satisfy regulatory audit requirements.
• Developing hybrid human-AI decision systems that orchestrate traders, risk officers, and models in real time.

This shift reflects a move toward systems thinking, ie designing adaptive, interdependent ecosystems rather than standalone tools. The core uncertainty is no longer ‘can it run?’ but ‘can it reason and remain compliant?’

Policy and tax: a new innovation framework

HMRC’s updated approach aligns fiscal incentives with the UK’s AI and Data Strategy. By recognising data infrastructure and mathematical research as eligible, the government is broadening the scope of innovation.

This benefits financial firms by allowing:

• Quant and risk teams can now capture the scientific effort spent formulating new quantitative methods, even when no prototype system yet exists.
• RegTech and compliance units developing algorithmic controls for SMCR or Basel IV reporting can evidence their R&D under the 'systematic, investigative' test.
• Data-centric projects (eg synthetic data generation, real-time entity resolution) can now claim qualifying expenditure, provided the technical uncertainties are defined and resolved methodically.

The focus is shifting from code deployment to knowledge advancement.

Strategic implications for financial leaders

For CTOs, innovation leads, and R&D claim owners, the message is clear: competitive advantage will come from designing intelligent systems, not just faster software.

Firms that integrate AI alignment, mathematical modelling, and human-in-the-loop design will be better positioned to:

• Demonstrate compliance-by-design under evolving FCA and Bank of England regulations.
• Use R&D tax relief to fund responsible AI development.
• Attract capital and talent by showcasing genuine technical innovation.

The future of R&D in financial services

The future of R&D in financial services lies in cognitive systems such as risk engines that reason, compliance tools that interpret, and trading models that explain themselves. Just as previous decades were defined by electronic trading and cloud migration, the 2020s will be shaped by systems that learn.

HMRC has opened the door. Now it’s up to financial institutions to step through and embrace systems-level innovation.

To discuss how we can support your R&D activity in financial services, please contact Constantine Costas.

Latest related insights