Aveni Unveils FinLLM: A Tailored AI Model for UK Financial Services

AI fintech firm Aveni is bringing generative AI to the UK financial sector with the launch of FinLLM, a large language model purpose-built for regulated financial services environments.

FinLLM was developed over the last six months at Aveni Labs, drawing on both advanced natural language processing expertise and a deep understanding of compliance and risk frameworks in financial services. Designed specifically for the UK market, the model aligns with emerging FCA guidelines and the EU AI Act, and is built with transparency, safety, and regulatory standards at its core.

What sets FinLLM apart is its financial specialism. Unlike general-purpose language models, FinLLM is tuned using financial data (both structured and unstructured) to support use cases that demand precision and trust. From regulatory compliance to customer service automation, the model is designed to enhance performance and to keep institutions within strict governance boundaries.

The project has been strongly supported by two of the UK’s major financial institutions, Lloyds Banking Group and Nationwide, both of whom invested in Aveni last year. Their involvement helped shape the direction of the model, ensuring it remains grounded in practical industry needs. Aveni’s collaborative approach brought together in-house AI experts and partner data teams to test and refine FinLLM on real-world applications.

Performance benchmarking shows FinLLM consistently outperforms more generic models on financial tasks and retains robust results on broader evaluation sets. This performance is enabled by a data strategy and evaluation process focused on continual pre-training and fine-tuning to ensure accuracy, explainability, and resilience.

Lloyds Banking Group’s Group Chief Data and Analytics Officer, Ranil Boteju, described FinLLM as a “game changer” for the sector, highlighting its potential to support a range of AI use cases across the bank. Nationwide’s Chief Data Officer, Sri Kanisapakkam, echoed that sentiment, highlighting the potential for the model to improve customer service through responsible innovation.

According to Aveni CEO Joseph Twigg, FinLLM reflects the growing importance of AI sovereignty and the UK’s ability to build its own technology infrastructure.

“As adoption accelerates, it’s critical that we retain the capability to develop and scale high-performance models within the UK,”.

Dr Lexi Birch, Head of Aveni Labs, spoke about the collaboration between academic researchers and financial professionals in shaping FinLLM’s roadmap.

“This model is built for a sector where accuracy and trust are non-negotiable. We’ve created something that reflects those priorities from the ground up”.

With its core model now released, Aveni will begin integrating FinLLM into its product suite, including Aveni Detect and Aveni Assist, supporting a wider shift toward domain-specific AI agents. Future development will focus on scaling the model across industry environments, enhancing its capabilities with proprietary datasets, and fine-tuning alignment strategies to meet the demands of live deployment.