AI in Finance Adoption Statistics
About 75% of UK financial services firms reported using AI in 2024, up from 58% in 2022; generative AI accounted for roughly 17% of those use cases. The data below comes from regulator surveys and published industry research, each datapoint with its source and year for traceability. None was measured by this site. The consistent pattern across sources: adoption is broad, generative AI is the accelerating edge, and governance frameworks are still catching up.
On This Page
Statistics
The numbers worth quoting
About 75% of UK financial services firms reported using AI in 2024, up from 58% in 2022
The joint regulator survey also found a further 10% of firms planning to adopt AI within three years, leaving a small minority with no AI use or plans.
Foundation models accounted for roughly 17% of all AI use cases reported by UK financial firms in 2024
Generative and foundation-model use cases were a minority of total deployments but the fastest-growing category, concentrated in earlier maturity stages than established predictive models.
Generative AI could add the equivalent of 200 to 340 billion US dollars in annual value to the global banking sector
The estimate corresponds to 9 to 15 percent of operating profit, with the largest gains projected in software engineering, customer operations, and marketing functions.
Around 55% of AI use cases involved some degree of automated decision-making, but only about 2% were fully autonomous
Most AI applications kept a human in the loop. Fully autonomous decision-making remained a small share, reflecting caution around accountability and explainability.
The top perceived AI risk among UK financial firms was data-related: data quality, privacy, and security
Third-party dependency and model complexity ranked among the next-highest concerns, signalling that the binding constraint is governance and data, not model capability.
The IMF warned that AI-driven trading can amplify market volatility and increase the speed and correlation of asset-price moves
The Fund flagged that wider AI adoption in trading could deepen flash-crash dynamics and herding, while also improving liquidity provision in normal conditions.
Roughly one third of firms cited a lack of internal AI skills or talent as a barrier to wider adoption
Skills shortages sat alongside data and governance constraints, indicating the bottleneck to scaling AI is organizational capacity as much as technology cost.
Key Takeaways
Methodology
Figures are compiled from regulator surveys and published industry research and reported with their original source and year. Where a survey reports a range or a year-over-year change, both endpoints are shown. No statistic on this page is derived from data collected by this site.
Try These Tools
Run the numbers next
Model Selector for Finance
Input task, latency budget, cost budget, context size, and quality sensitivity; get ranked model recommendations with rationale — grounded in published.
Hallucination Detector
Paste a source document + an LLM's extraction. Every numeric claim in the output is checked against the source. Client-side. Catches silent fabrication.
Finance MCP Directory
Security-graded catalog of finance MCP servers — Alpaca, Polygon, Databento, IBKR, Tradier, Tiingo, NautilusTrader. Scope, auth, idempotency, transport.
Sources & References
- Artificial intelligence in UK financial services 2024 — Bank of England and Financial Conduct Authority (2024)
- The economic potential of generative AI: The next productivity frontier — McKinsey & Company (2023)
- Global Financial Stability Report, October 2024 — International Monetary Fund (2024)
Related Content
Keep the topic connected
MCP (Model Context Protocol)
Model Context Protocol: Anthropic's open standard for letting LLMs discover and call tools — the interface, why it matters, and finance MCP server checks.
Hallucination Detection
Detecting LLM hallucinations in financial outputs: the verifiable-claim approach, citation grounding, and cross-model agreement signals that work.
Regulatory Cost of AI in Finance
Regulatory cost as a function of jurisdiction, model class, and end-use: the FTC vs NLT distinction and the documentation burden by regime.