• Nearly three-quarters (70 per cent) of FS firms surveyed said they are at the pilot stage for generative (content production-focused) AI using ‘co-pilot’ type employee efficiency tools, but it is expected that return on investment will take between three and five years for more advanced use cases.
• Three-quarters (75 per cent) of FS firms believe they will benefit from generative AI – with the biggest opportunities expected from productivity improvement and operational effectiveness rather than customer-facing, revenue-related, use cases.
• When Marsh McLennan tested a generative AI assistant with staff, 94 per cent of users said it improved their productivity.
• 95 per cent of firms surveyed are already accounting for AI risks in their control frameworks, and three out of five (60 per cent) have already taken action to prepare for the risks of generative AI.
• Four out of five FS firms (80 per cent) say collaboration with regulators is beneficial to highlight best practices in AI and create an internationally aligned regulatory system.

Jana Mackintosh, Managing Director of Payments, Innovation and Resilience at UK Finance, said:

Generative AI is a new horizon for financial services, with clear opportunities for firms, from boosting productivity and growth, to improving the experience for customers. And as this report shows, a majority are already piloting the technology to find the best way to safely explore these exciting opportunities.

“At the same time, however, FS firms are clear-eyed about the risks. The sector has been using predictive AI for years, and firms have enhanced their risk management processes to make sure the technology only ever adds to their accuracy, security and offering for customers. And the majority are already upgrading these processes to lock down the risks from generative AI.

“What firms want now is to collaborate with policy makers and regulators on a long-term, flexible approach to regulation that can keep pace with technological change, ensuring our FS sector can guard against any risks from this technology, while safely seizing its great competitive advantages.

Sian Townson, Partner, Oliver Wyman said:

With the financial services sector in the throes of technological change, AI has the power to turbocharge opportunities, if used correctly. Adoption should be balanced with nascent regulation, both within the UK and internationally.

“Our survey shows that while adoption is widespread in the industry, firms are clear about the risks to ensure best practice to meet both customer needs but also wider competitive advantages that AI can bring.

Nearly three-quarters (70 per cent) of financial services firms are in the pilot stage of testing the use of generative (content production-focused) AI, while nine out of ten (91 per cent) already use predictive AI for functions including fraud detection and risk analysis.

Three-quarters of firms believe generative AI will benefit them to at least the same level as predictive AI, with the main expected benefits in process automation, sales and customer service functions. Survey respondents said the top three expected benefits in these areas are better productivity (74 per cent), operational effectiveness (70 per cent) and better customer experience (61 per cent).

This is borne out by the case study of Marsh McLennan, which deployed its own generative AI assistant called LenAI. It was used by 15,000 staff within 30 days of launch, of whom 94 per cent said it improved their productivity and efficiency. The main uses were writing and improving communication (70 per cent), searching for answers to specific questions (68 per cent) and summarising documents (56 per cent).

Firms are also clear, however, about the risks posed by this new technology, and three out of five say they are already adapting their existing risk management frameworks to include generative AI.

To safely seize the opportunities from generative AI, four out of five firms (80 per cent) want to collaborate with regulators to define best practice. Firms also want to ensure international alignment of regulations, with 65 per cent saying conflicting rules between jurisdictions is one of their top-three concerns about AI policy. The report also includes a range of possible statutory and non-statutory options to tackle the policy challenges currently facing the sector.