Bridging the skills gap in financial services through ‘fake’ data
1 OCT, 2025
By Tracy Vegro from Chartered Institute for Securities & Investment
by Tracy Vegro OBE, CEO Chartered Institute for Securities & Investment
Tonic.ai is widely-seen as the poster child of what it calls “fake data” – the synthetic data on which the AI industry and those it serves increasingly rely to deepen and broaden its reach. It was born from a tangible, practical need: to equip developers with high-quality, realistic data for development and testing. The firm grew out of Palantir, a US-based software mammoth that builds data integration and analytics platforms - most notably Foundry and Gotham - widely used by financial institutions and governments, where it has deep engagement in military and intelligence operations to support defence, surveillance, and battlefield decision-making.
Tonic co-CEO Ian Coe recalls the spark, from a common problem he faced at Palantir: “If you were having an issue on site, you couldn’t just send the data over the wire back to Palo Alto where somebody, maybe a developer, was sitting.” This apparently simple challenge – the inability to share sensitive data for troubleshooting – was the catalyst for Tonic’s launch.
In finance, synthetic data is playing a pivotal role in addressing the critical global shortage of artificial intelligence (AI) and machine learning (ML) skills in the sector.
AI is forecast to transform our industry over the next five years, with an estimated contribution in the UK alone of £35 billion and up to a 50% increase in productivity. However, the sector is facing a significant shortfall in professionals trained in AI and ML - one that threatens to limit the UK’s and most other countries’ competitiveness and capacity for innovation. The pace of technological advancement has outstripped the ability of traditional education and training providers to keep up.
Synthetic data yields a practical, scalable solution: the creation of synthetic data ‘lakes’, ideally developed hand-in-hand with regulators. In UK, the Financial Conduct Authority (FCA) is already advanced in planning this for its enhanced digital sandbox. This environment will provide course developers - ranging from professional bodies and business schools to apprenticeship schemes and commercial training providers - with access to realistic, privacy-preserving datasets that simulate financial data. These datasets can be used to build immersive, hands-on AI training programmes specifically designed for financial services applications.
Synthetic data is essential in regulated industries because it allows AI systems to train and learn from realistic scenarios without exposing sensitive personal, financial, or health information, thereby ensuring compliance with strict privacy and data-protection rules. It also enables richer, more diverse, and edge-case experiences than real datasets alone can provide, improving model robustness and safety in high-stakes environments.
Scaling access to synthetic data tools enables industry to innovate safely, regulators to test and oversee compliance more effectively, and education providers to train future talent with realistic, risk-free datasets. The policy implications and opportunities for public-private collaboration lie in creating shared standards and secure frameworks that enable innovation with synthetic data while safeguarding privacy, fostering trust, and accelerating responsible adoption across sectors.
Sharp eyes in Britain then are on the FCA’s Digital Sandbox and the work it hosts helping accelerate digital talent development.
