Mary Emma Barton Research and Analysis, Financial and Monetary Systems, World Economic Forum
Bryan Zheng Zhang Executive Director, Cambridge Centre for Alternative Finance, the University of Cambridge Judge Business School
- Global AI in Financial Services Survey, supported by EY and Invesco, shows the impact AI will have on financial institutions, from business models to changes in the workforce;
- By 2030, FinTechs anticipate AI will have expanded their workforce by 19%;
- Data quality and access to data, as well as access to suitable talent, are all seen as major obstacles to implementing AI.
Artificial intelligence (AI) is in the process of transforming a variety of models in the global financial services industry, a global survey jointly conducted by the Cambridge Centre for Alternative Finance (CCAF) at the University of Cambridge Judge Business School and the World Economic Forum suggests.
The study, supported by EY and Invesco, demonstrates that AI is changing how financial institutions generate and utilize insights from data, which in turn propels new forms of business model innovation, reshapes competitive environments and workforces, engenders new risk dynamics and poses novel challenges to firms and policy-makers alike.
Advancing towards mass adoption
The survey, which gleaned responses from 151 financial institutions, including both incumbent firms and FinTechs hailing from more than 30 countries, confirms AI as a crucial business driver across the industry in the short term. Notably, AI adopters do not appear to have specific modi operandi for implementing AI; instead, 64% expect to become mass adopters within two years, proving the growing potential of AI to stimulate innovation and growth across a wide range of business functions.
FinTechs and incumbents alike are moving from mainly using AI to reduce costs to utilizing its capabilities for revenue generation, albeit pursuing different AI strategies to achieve this. Most incumbents primarily use AI to enhance existing products and services, whereas many FinTechs use it to create new value propositions, as shown in the chart below.
This strategy is concomitant with selling AI as-a-service, with 45% of all FinTechs (excluding B2B-only companies) offering AI-based B2B solutions compared to only 21% of incumbents.
EY’s Global AI Leader, Nigel Duffy recognizes the importance of understanding the implications of mass adoption: “AI is transforming the Financial Services industry and we can expect widespread adoption to continue. As the technologies give way to new revenue streams and transform business functions, it’s increasingly important for organizations to focus on the long-term implications of AI adoption.”
In the wake of mass adoption, survey participants’ perceptions indicate that AI may replace nearly 9% of incumbent financial services jobs by 2030, while FinTechs anticipate AI will expand their workforce by 19% in the same time frame. Reductions are expected to be highest in investment management, with participants anticipating a net decrease of 10% within five years and 24% within 10 years.
New models, new challenges
In the race to adoption, companies face similar hurdles. Data quality and access to data, as well as access to suitable talent, are all seen as major obstacles to implementing AI by more than 80% of respondents. The cost of hardware/software, market uncertainty and technological maturity appear to represent lesser hindrances.
Yet, even when these implementation hurdles are overcome, the proliferation of AI poses a range of challenges for all parties involved in the financial services landscape:
- AI implementation is expected to lead to an exacerbation of certain market-wide risks and biases. For instance, firms expect AI to create or exacerbate bias in credit analytics, especially when non-traditional datasets are used;
- While views of regulatory influence on AI implementation diverge, most firms feel impeded by data-sharing regulations between jurisdictions and entities as well as regulatory uncertainty and complexity;
- Nearly half of all respondents see Big Tech firms, such as Google or Tencent, using AI capabilities to enter the financial services market as a major competitive threat.
Apart from underpinning these findings with empirical quantitative data, the study also identifies strategy-related aspects which can be generalized across different sectors and entity types. Indeed, synthesizing survey results allows for the conclusion that any firm seeking to develop a successful AI strategy will need to secure sustainable and (ideally) exclusive sources of training data. While underlying algorithms and systems may be complex, they are amenable to commoditization and represent a lesser differentiator than unique datasets.
The universal need for data at scale encourages the creation of digital platform models which integrate AI-enabled products and services, forming data-rich interfaces between buyers and suppliers. This is already visible in critical tech sector players such as Google who have taken advantage of the self-reinforcing characteristic of AI at scale to establish dominance in search.
It remains unclear, however, in which direction the power dynamic between incumbents, FinTechs and Big Tech will evolve, especially given the complementary capabilities they bring to the table. Being able to gain a comprehensive overview of these overarching developments will require further in-depth research on the mechanics of early adopter advantages in AI, the burden of legacy infrastructure for incumbents, AI-empowered network effects and AI-induced biases and risks.
Despite these challenges, Invesco’s Chief Technology Officer, Donie Lochan, notes the incredible opportunities AI creates for financial services: “The report highlights the amazing opportunity ahead of us in financial services for using artificial intelligence and machine learning to the benefit of our customers and our organizations. Technological advances such as leveraging intelligence to define investments for customers tied to their personalized goals, improving customer experience through the use of intelligent bots, additional alpha generation via insights from alternative datasets, and operational efficiencies through machine learning automation, will soon become the norm for our industry.”