The Trillion-Dollar Question: Capturing ROI when 60% of Financial Institutions Report Little Profit Impact from AI
Almost daily, the market is presented with another bold artificial intelligence announcement or prediction: a targeted $2 trillion valuation for the IPO of Anthropic, the AI company behind the Claude models; the ever-increasing share price of AI chip manufacturer Nvidia, which has risen more than 1100% since the launch of ChatGPT to the public in 2022; and the torrent of AI-generated content now estimated to make up around half of newly published English-language articles online, all building the case that generative AI has the potential to fundamentally transform the way we work and live our lives.
However, for all the noise, there is still little conclusive proof that AI adoption is driving real profitability and revenue growth in the global economy, while AI-related investment is expected to reach $1 trillion this year according to Goldman Sachs.
For the financial sector, a survey published this year by the Cambridge Centre for Alternative Finance (CCAF), covering fintechs, traditional financial institutions and regulators, found 60% of respondents either could not identify or found no overall profitability through revenue growth from AI adoption.
Rather, the real benefit, at least for now, will seemingly come from improved productivity.
The CCAF study found that financial services firms expected the greatest value drivers to be cost reduction and resource optimisation (57%) as well as speed, scalability and latency (53%), while only 24% of respondents saw increased profitability and revenue growth as a key AI value driver.
Trust the Financial Services Track Record
For large financial institutions, the productivity value from AI is significant especially considering that AI usage among finance workers is the second highest of any industry, behind only technology.
However, impactful AI adoption and integration across an enterprise rather than just at an individual’s desktop is another challenge entirely.
For businesses, this is a massive stumbling block with MIT highlighting a staggering 95% failure rate for enterprise AI solutions – ultimately capturing no value as they fail to move from a proof of concept to an integrated, business-accepted solution.
While generative AI vendor solutions provide tools that supercharge the velocity of coding, data processing and analysis, and writing, the scalability of an AI-enabled solution is being hindered by unreliable, poorly controlled and ultimately generic AI solutions that do not solve real business issues.
In financial services, the highest-value AI applications are usually in internal operations rather than at the customer interface.
Back-office processes are the best fit for AI automation because they are high-volume, rules-based and data-intensive.
Common use cases include:
- Intelligent document processing to replace manual data capture
- AI-generated commentary and insights for management and regulatory reporting
- AI-assisted coding to speed up legacy system migration and automated testing
- AI-assisted reconciliation and exception handling across large datasets
- AI-powered real-time monitoring of fraud and financial crime
However, financial services firms are well positioned to unlock the value that comes from scaling AI solutions, particularly considering the significant investment they have made over the last few years in their data and technology capabilities including single view of customer, cloud infrastructure migration and advanced analytics.
Furthermore, many of the large banks and insurers in South Africa have established and are maturing their AI frameworks that set out key practices and policies related to AI use case selection, AI lifecycle management as well as risk and data management in relation to AI.
Capturing Productivity Gains
Economies are shifting from a period of experimentation and exploration towards long-term application of artificial intelligence, but gains are expected to be uneven.
The World Bank predicts that expected productivity gains in advanced economies will be double that of emerging markets like South Africa, due to their higher usage and exposure rates to AI.
This scale of South Africa’s limited exposure is clear to see when a nationally representative survey by the Human Sciences Research Council (HSRC) and the Global Center on AI Governance found that more than a third of South Africans have never even heard of artificial intelligence (37%), while a further 36% indicated they knew little about AI, let alone made use of it.
For financial services, it will be critical that time saved through AI-enabled efficiency is redeployed productively, and that employees are directed to high-value tasks while reviewing and ensuring high-quality outputs from AI solutions.
While it is still unclear whether AI will lead to substantial job losses, the issue of reskilling and enabling employees to build credible AI-related skills is becoming not only an imperative for businesses but an imperative for every economy.
In response to the rapidly evolving state of artificial intelligence in society and the working world, Monocle has launched its inaugural AI Sentiment Survey to understand how people really feel about AI today in both their workplace and personal lives.
The survey is anonymous and will take five minutes.
The findings will be used to identify key trends and perspectives and contribute to the broader conversation around AI and its evolving role in our lives and work.
Loading ...