Harnessing gen AI in financial services: Why pioneers lead the way
A Deloitte survey finds financial services organizations with high confidence in their generative AI expertise also report greater rewards from their gen AI initiatives
2025 could be a pivotal year for the financial services industry, as it could begin to reap more benefits from generative AI, transitioning from experimental pilots to implementation. However, the level of preparedness and pace of adoption among institutions vary significantly. Our analysis of Deloitte’s State of Generative AI in the Enterprise survey, focusing on about 540 respondents in the financial services industry, reveals a clear separation among senior leaders and their attitudes toward gen AI: pioneers, or early adopters, who rate their organizations as having very high or high expertise in generative AI, and followers, who have some, little, or no expertise with the technology. Our findings indicate that 46% of respondents possess the character traits of pioneers (see methodology).
That said, survey responses over four iterations between 2023 and 2024 reveal a notable shift in self-assessed expertise based on organization size. Initially, in late 2023, smaller firms (with annual revenues between US$500 million and US$5 billion) were evenly split between high and low expertise, whereas 59% of larger firms (with annual revenues over US$5 billion) rated themselves as having low expertise (or followers, as we classify them). However, recent data shows that larger firms have made progress and half of them now identify as pioneers, indicating that as organizations continue to engage with technology, their self-assessment of expertise is likely to evolve, with some building confidence in their skills and others realizing they have more work to do.
Early adopters, based on their responses, exhibit greater interest and trust in the technology. They are three times more likely than followers to believe that generative AI can significantly transform their organization within a year. This belief is likely driving their investment decisions: Seventy-six percent of pioneers allocate 20% or more of their artificial intelligence budgets to generative AI, compared with just 46% of followers. In the first iteration of the survey, conducted between October and December 2023, the figure stood at 62% for pioneers and 38% for followers.
While this leap of faith could potentially distinguish pioneers in the future, is it currently generating the returns this group is expecting? If so, where are these returns occurring?
Pioneers unlock substantial value from gen AI deployments
Pioneers acknowledge seeing significant value from their gen AI deployments thus far. In the previous iteration of the survey, fielded between May and June 2024, 78% of pioneers attributed the rise in investment to the strong value they’ve observed from their generative AI initiatives. In the latest survey, conducted between July and September 2024, an additional question was included, asking respondents to estimate the return on investment (ROI) achieved from their most advanced gen AI initiative.
The shaded region in figure 1 below illustrates the estimated additional value that pioneers have realized from this initiative. The gap between the pioneers’ and followers’ estimates of expected ROI is noticeable, with 74% of pioneers estimating ROI of more than 10%, compared with 44% of followers. Interestingly, this gap is wider for smaller financial services organizations, indicating that followers in this category are likely facing considerable challenges in deriving value from their gen AI efforts. However, new, more cost-effective gen AI models may lead organizations to revisit their ROI expectations.1
Nearly half of the pioneers (47%) surveyed estimate that ROI from their advanced generative AI initiatives exceeds their expectations, compared with only 17% of the followers. They are twice as likely as followers to acknowledge that they have achieved their desired benefits to a very large or large extent. The benefits cited by pioneers, in order from most to least mentioned, include uncovering new ideas and insights, driving innovation and growth, improving efficiency and productivity, detecting fraud, enhancing existing products and services, and strengthening client relationships.
Pioneers spearhead scaling gen AI adoption within their organizations
Among pioneers in our survey, 43% report that they’ve provided access to gen AI tools and applications to over 40% of their workforce. Just 19% of followers say they’ve done the same. Additionally, a significant number of pioneers report they’ve moved beyond an experimental pilot stage to implementation, with limited or at-scale implementation in areas such as IT and cybersecurity; marketing, sales, and customer service; and strategy and operations (figure 2).
While we do not know how the following organizations would rate their generative AI expertise, there are increasing examples of the technology’s usage in the marketplace. For instance, in June 2024, Morgan Stanley announced the launch of a gen AI tool, AI @ Morgan Stanley Debrief, which is capable of summarizing video meetings and generating follow-up emails for financial advisers.2 Ally Financial is using the technology to create marketing campaigns and content.3 Meanwhile, insurers and reinsurers, such as Prudential, Munich Re, and AIG, are increasingly using gen AI for underwriting and claims processing.4
Followers, particularly those from smaller organizations, on the other hand, may be waiting to see more substantial value. After an initial rush of confidence, the experience they gained is perhaps contributing to a more measured pace, as many are still focused on experimenting and piloting use cases (figure 3). Data suggests that pioneers have so far realized the most significant value in IT and cybersecurity, as well as product development functions. This success likely explains why this group has progressed beyond the experimentation phase. The gaps in experimentation and implementation between the two groups, as seen in the charts, are the widest in these areas. For example, the cybersecurity function of a bank, as highlighted in Deloitte’s recent gen AI report, could sift through millions of incoming cyber-threat alerts to identify and prioritize “real threats” using a gen AI–based solution.5
Despite the technology’s potential, both pioneers and followers have yet to achieve significant breakthroughs in the areas of human resources and legal, risk, and compliance. In the latter case, gen AI can help users ask fact-based questions, compare and extract requirements from multiple regulations, assess compliance gaps by comparing regulations to internal policies, and update and communicate changes to policies and procedures through intuitive interfaces.6 Meanwhile, the most prevalent use case for the technology in human resources is sourcing and recruiting, as highlighted by participants in an online executive forum.7 However, aside from talent acquisition, it is possible that the perceived value in these areas does not yet justify the cost, even for pioneers.
Preparation is the key to unlocking value
When asked about their readiness to broadly adopt gen AI tools and applications, only 7% of followers surveyed feel highly to very highly prepared in terms of talent, 16% in risk and governance, 20% in technology infrastructure, and approximately 30% in strategy and data management. Meanwhile, the majority of pioneer respondents feel highly prepared, except in the areas of talent (37%) and risk management (36%). Furthermore, twice as many followers as pioneers surveyed cite the lack of an adoption strategy and executive commitment as barriers to gen AI adoption.
Perhaps for these reasons, only 22% of followers said they plan to fully scale 40% or more of the projects currently in the experimentation or proof-of-concept stage over the next three to six months, compared with nearly half of the pioneers.
Pioneers’ readiness in the key areas of talent, risk and governance, data management, and adoption strategy is positioning them to scale their generative AI efforts and create short-term value. But gaining a sustained competitive advantage could ultimately hinge on applying the technology more strategically to bolster one’s existing strengths. Therefore, it remains to be seen whether pioneers or followers will come out ahead in this race.
Methodology
The statistics in this article and its graphics are drawn from Deloitte’s State of Generative AI in the Enterprise series. Overall, 2,773 leaders from 14 countries participated in the survey, which was conducted between July and September 2024. Among these participants, a subset of 542 financial services leaders provided the insights prominently featured in this article. Respondents were senior leaders within their organizations, including board members, C-suite executives, as well as those holding president, vice president, and director positions.
Prior iterations of the State of Generative AI research surveyed 2,770 leaders between May and June 2024, of which approximately 518 were from financial services; 1,982 leaders between January and February 2024, including around 357 from financial services; and 2,835 leaders between October and December 2023, with about 513 respondents from financial services. Due to rounding, percentages in this report may not total 100%.
Pioneers are defined as respondents who rate their organizations as having very high or high expertise in generative AI. Followers are defined as respondents who have some, little, or no expertise in the technology.
Generative AI is an area of artificial intelligence and refers to AI that, in response to a query, can create text, images, videos, and other assets. Gen AI systems can interact with humans and are often built using large language models.
