By Roy Urrico
Finopotamus aims to highlight white papers, surveys and reports that provide a glimpse as to what is taking place and/or impacting credit unions and other organizations in the financial services industry.
Santa Clara, Calif.-based technology company NVIDIA unveiled the top four findings garnered from its State of AI in Financial Services: 2023 Trends survey taken by nearly 500 global financial services professionals. This year’s study revealed for banking institutions, and fintech, insurance and asset management firms, the goals remain the same — find ways to accurately manage risk, enhance efficiencies to reduce operating costs, and improve experiences for clients and customers.
NVDIA, which works with dozens of leading banks to develop and execute rapidly evolving AI strategies and also designs and manufacturing graphics processing units (GPUs), also disclosed benchmark test results for its NVIDIA A100 Tensor Core GPUs in a separate announcement.
NVIDIA Vice President of Financial Services, Malcolm deMayo provided Finopotamus with a deeper dive on the survey and benchmark results.
The Final AI Four
Nvidia revealed the top four findings in the State of AI in Financial Services: 2023 Trends:
· Hybrid cloud is coming on strong. This year’s survey found almost half of respondents’ firms moving to the hybrid cloud to optimize AI performance and reduce costs. DeMayo noted that hybrid cloud or multi-cloud use is growing in a strong way. “Hybrid cloud is the use of a public cloud and a private cloud on premise. So that data that needs to live in a safe, secure, private place, can make regulators happy. (Data) stays where it needs to stay.”
● Large language models top AI use cases. The survey found the most popular AI uses as natural language processing and large language models (26%), recommender systems and next-best action (23%), portfolio optimization (23%) and fraud detection (22%). The metaverse/virtual worlds attained 12% penetration among respondents, as financial services companies identify a range of use cases including training, new employee onboarding, retail branch simulation, and insurance risk evaluations.
“AI is used to improve customer service in financial services firms, leveraging virtual assistance or digital humans to improve customers’ experiences,” explained deMayo. “These AI technologies can do things like answer questions, translate voice to text, text to voice, and they can answer questions quickly.”
● Banks seeing more potential for AI to grow revenue. Nearly half of survey takers said that AI will help increase annual revenue for their organization by at least 10%. More than a third noted that AI will also help decrease annual costs by at least 10%.
Financial services professionals highlighted how AI has enhanced business operations — particularly improving customer experience (46%), creating operational efficiencies (35%) and reducing total cost of ownership (20%). AI also helps prevent fraud by enhancing anti-money laundering and know-your-customer processes, while recommenders create personalized digital experiences for a firm’s customers or clients.
“The industry is starting to recognize that AI is a gold mine, not just driving efficiency, but also for being smarter about how you approach your customers and how you interact with the markets,” said deMayo. He also pointed out that management consulting firm McKinsey & Company in a report believes AI can potentially unlock $1 trillion of incremental value annually in financial services.
● The biggest obstacle: recruiting and retaining AI talent. Recruiting and retaining AI experts is a problem reported by 36% of respondents. There is also inadequate technology to enable AI innovation, according to 28% of survey takers. Another pressing issue observed by 26% of financial services professionals is insufficient data sizes for model training and accuracy. Nvidia suggested the use of generative AI to produce accurate synthetic financial data used to train AI models.
DeMayo described an intense competition to recruit data scientists and quants (quantitative analysts) between financial firms “and obviously with the tech firms too.”
Recently Deutsche Bank announced a partnership with NVIDIA to accelerate the use of AI and machine learning in the financial services sector. Initiatives include exploring how to engage employees, potential recruits and customers more interactively, improving experiences using 3D virtual avatars in real time, 24/7. DeMayo explained Deutsche Bank also plans to “partner with universities to build their AI capability set so that the smartest, the brightest and the best want to come work for them.”
Executive Support for AI at New High
Financial institutions plan to continue building out enterprise AI in the future. This will include scaling up and scaling out AI infrastructure, including hardware, software and services.
Nvidia’s survey also found increasing executive buy-in for AI is a new theme in the survey results. Some 64% of those surveyed noted that “my executive leadership team values and believes in AI,” compared with 36% a year ago. In addition, 58% said that “AI is important to my company’s future success,” up from 39% a year ago.
DeMayo said “We're having conversations on a regular basis with our partners in ecosystems, the OEMs (original equipment manufacturers) that we work with (Dell, Lenovo and others) and the cloud service providers. We actually do not sell anything direct. Our charter is to help, advise and learn from our customers what is going on.”
Credit unions would engage through Nvidia’s partner ecosystem. “We have, trained over 4,000 ISVs (independent software vendors). We have trained tens of thousands of data scientists and developers across the world on how to use accelerated computing and Nvidia AI and NVIDIA Enterprise more efficiently and more effectively in whatever they are trying to get done,” deMayo stated.
Benchmarking Machine Learning
NVIDIA recently announced the results of the highly technical measurement of its A100 Tensor Core GPUs running on its Supermicro servers in the latest STAC-ML (Securities Technology Analysis Center-Machine Learning) inference benchmark. The study analyzes the latency of long short-term memory (LSTM) model inference, which is the computing time from receiving new input data until the model output.
Financial institutions closely followed the results since three-quarters of them rely on machine learning, deep learning or high performance computing, according to Nvidia’s State of AI in Financial Services survey. LSTM is a key model approach used to discover items such as financial time-series data like asset prices.
The benchmark includes three LSTM models of increasing complexity. NVIDIA A100 GPUs, running in a Supermicro Ultra SuperServer, demonstrated low latencies in the 99th percentile. Said deMayo, “Predictable performance is crucial for low-latency environments in finance, as extreme outliers can cause substantial losses during fast market moves.”
DeMayo illuminated how latency — the roundtrip time it takes from input to output —shrinks the time window and helps firms develop a competitive edge over their peers. “There's a convergence of algorithmic systems happening on GPU accelerated computing. The volume, the velocity and the variety of data — of new data — created has changed the game. It is surpassing the capabilities of (traditional) CPUs (central processing units).”
The use of LSTM models is an AI enabler. Explained deMayo, “It is proving that GPUs can replace lower level hardware devices, that are used today like FPGA (field programmable gate array) and ASICs (application specific integrated chips). These are very expensive and hard to work with technologies. It extends the use of a general purpose GPU, which is a good thing because you drive down costs.”
DeMayo added financial institutions such as credit unions, banks, and trading firms, do not have unlimited resources to spend, and they are looking for belt tightening ideas, around AI adoption. “One of the cool things about this (benchmark) announcement is it allows financial institutions’ quant research and risk management groups on the same platform to standardize, consolidate and harvest some cost savings. It is a unique advantage of NVIDIA's GPU accelerated computing.”
He added, Nvidia has helped top banks. “In particular one in New York City has reduced their trade computational grid farms’ total cost of ownership by 80%.” While also allowing that financial institution to run more simulations. “We think this is yet another important validation that GPUs can augment and help CPUs deal with the volume, the velocity, and the variety of data that banking and financial firms have to deal with today.”