While much has been written about the promise of Artificial Intelligence (AI) in Financial Services, there is noticeably less literature on how firms can successfully implement and gain value from these technologies. Our view is that to realise success, organisations need to rebalance their existing strategic risk and governance frameworks against a new set of demands presented by AI.

In late 2025, we surveyed a selection of 34 Financial Services organisations to investigate their attitudes towards long-term AI implementation, governance, and strategy. The respondents represent a cross-section of sectors, revenues, functions, and seniority levels – offering an expansive and illustrative perspective of this emerging subject.

For the full report including a section on the risks of AI and our conclusions, click here. 

Summary of findings

Most organisations are in the early stages of AI implementation. Our survey reveals that only 22% of respondents have one or more applications in wide use, excluding tools like ChatGPT or Copilot, while 78% have yet to implement any AI technology and only half are in the consideration or planning phase. This suggests that AI usage in Financial Services may be less widespread than anticipated. While this data should not encourage complacency, it reassures firms that they can proceed in a measured way, rather than feeling that they need to rush to keep up with the hype.

Encouragingly, some firms are demonstrating strategic thinking. Principles for AI deployment, such as fairness, sustainability, accountability, and transparency, are acknowledged as important by 50% of respondents. However, many organisations are still struggling to establish a meaningful strategic direction for AI. Only one respondent reported having a well-embedded AI strategy, while 23% have a strategy that is unclear, and 27% have no AI strategy at all. Even if firms have not yet strategically incorporated AI it is crucial for firms to ensure a strategic direction for data governance, given the heightened data risks. Despite this only 17% have a fully embedded data governance strategy.

While firms might procced with implementing AI applications at a measured pace, there is urgency in developing and embedding a well-informed strategy. Without this, firms may be leaving success to chance and exposing themselves to significant impacts. AI introduces novel risks that require identification, analysis, and management, demanding critical thinking. Despite this, only 27% have assessed AI risk extensively, and alarmingly, 38% report that AI is not embedded in their risk management framework. With the market increasingly inclined towards third-party supply, it is concerning that over 70% have made no special provision for managing third-party risk.

A more optimistic outlook is suggested by some attitudinal indicators. For instance, 65% of respondents said their attitude to risk would be adapted to AI. They identified the main risks as unintentional customer harm, regulatory and legal breaches, and failure to realise benefits. Within 'customer harm,' discrimination, treatment of vulnerable customers, and accuracy of customer data were acknowledged as important factors. This picture is somewhat inconsistent with encouraging attitudes not manifesting in practical steps. As with overall strategy, firms need to develop their approach to understanding, managing, and remediating AI risks through thorough assessments and robust controls.

Further to attitudes to risk, 74% see more upside to AI than downside. While firms may be positively inclined, 56% classify their AI knowledge and capability as inadequate, which could hinder the ability to mobilise new technology. However, firms are proactively addressing this limitation. An encouraging 50% are scoping their needs, and 24% are investing in training. Only 9% expect AI capability to be recruited externally. Positive attitudes towards AI, coupled with recognition of its emerging nature and skill scarcity, are likely to lead more firms to invest in developing their own personnel.

A significant 94% of respondents understand their responsibilities regarding AI, with 82% confident in fulfilling them. With a respondent profile of 62% Board and C-suite members, this bodes well for firms' governance and management. However, only 38% have a specific individual responsible for AI, 32% have AI-specific policies and procedures, and 26% have added specific subject matter expertise into their governance structure. As with risk, this offers a different perspective on how firms are translating positive intentions into practical governance.

Expectations of success include 50% of respondents seeking higher accuracy rates, 41% aiming for process efficiency, and 32% targeting cost reduction. User satisfaction is also a measure of success, indicating a human component in what many fear may displace people. However, 70% have not seen any AI-specific management information (MI) to measure these successes, and 47% find usual decision-making cycles too slow.

If you would like to explore the report findings further, please contact BDO Digital Partner, Catherine Wilks.
________________________________________
Access the full report here
For the full report including a section on the risks of AI and our conclusions, click here
________________________________________