Nathan Gee, Global Head of Marketing, SmartStream
Afternoon sessions at SmartStream’s recent Connect EMEA 2024 event included an interesting discussion, moderated by Nick Smith, EVP of Customer Experience at SmartStream, about the impact of the cloud and AI on the financial services sector. Participants included Brian Hayes, Head of FSI GTM EMEA, AWS, Brendan Reilly, Global Head of Commercial Product, Moody’s, and Martin Lawrence, Chief Customer Officer, The Value Exchange. The participants offered their thoughts as to how AI and the cloud are reshaping the financial industry, as well as influencing innovation at their own organisations.
Regulation
The panel discussion initially turned to the question of regulation and its probable effect on AI deployment in the cloud. Brian Hayes noted DORA (the EU’s Digital Operational Resilience Act, which will take effect in January 2025, bringing major implications for financial institutions and their critical third-party technology service providers) but also pointed out that other pieces of emerging or re-emphasised regulation on AI were set to have an impact. Regulators, the panel observed, were beginning to understand that AI would be integrated with everything we do in the future.
How is AI transforming the financial industry?
SmartStream’s Nick Smith gauged the panel’s views on the transformative influence of AI. Perhaps surprisingly, Martin Lawrence of The Value Exchange argued it was only the retail sector that had really felt the change thus far, and not the institutional side. AI, Lawrence added, is currently being layered on top of existing technology but is not actually making money for firms. Highlighting this point, he referenced research carried out by The Value Exchange in which 65% of survey respondents from financial institutions saw AI as having a significant impact on costs, 55% a significant effect on risk, while 30% believed it impacted revenue. Some 25% said it had no impact on revenue and only 8% viewed it as a key investment priority. At present, the main rationale for deploying AI appears to be increasing individual productivity and enhancing existing processes.
The panel noted that specialisation is an area in which AI is likely to play a key role, for example, in the customisation of funds. Additionally, with T+1 on the horizon, heralding a compressed settlement cycle, AI-enabled predictive trade analytics – which can help firms to identify the chance of specific client trades failing – may prove useful. More generally, where risk management is concerned, financial institutions are still hampered by their siloed structure and require better informed decision-making. Here, AI can play a beneficial role, giving improved and speedier access to information, and promoting a proactive rather than a reactive approach.
Keys to successful AI implementation
The discussion pivoted to participants’ own experience of AI implementations and lessons that could be drawn from them. Brendan Reilly, reflecting on the development of Moody’s AI-based ‘research assistant’, underlined the need for deliberate and careful implementation, plus the importance of thinking differently about information. When using AI more generally in an organisation great caution is required about which information is exposed to the AI, he added, such as extra human controls around sign-offs on which SharePoint sites could be exposed to the AI, to ensure that, for example, only approved, and not draft documents were released into the AI.
Creating value for both the organisation and its customers
Brendan Reilly noted that Moody’s AI-based Research Assistant, an AI layer assisting access to the company’s research and ratings data, is beneficial for customers and Moody’s own analysts. The fact that is employed as a tool for internal use, as well as for customers, believed Reilly, helps development to be even more effective. The Research Assistant is a straightforward concept but has proved very powerful, with estimates showing that it can save up to 27% of Financial Analysts time conducting research and analysis,.
Data is key
Data is key, as is the sourcing of information, underlined Brian Hayes of AWS, and you must get it right if you want to avoid bias or hallucinations. Careful development of guardrails is highly important, too. Building an AI capability inside the organisation is the best way to start, building out the required capabilities in areas such as governance, as is taking small steps and continously learning and seeking to innovate, he suggested. Companies should also spend time starting to reimagine business processes. Interestingly, he noted that AWS does not view AI as a tool or feature but as a truly revoluntinary capability and approach to doing business.
Using AI to represent data through graphics
The panel also focused on the value of having an AI-based system that is able not only to make information easier to navigate but facilitates understanding through the generation of graphics. SmartStream’s Nick Smith pointed out the value of setting up AI to create visual representations of information to enhance users’ comprehension of data.
