[ad_1]
This article is part of a VB Lab Insights series on AI sponsored by Microsoft and Nvidia.
Don’t miss additional articles in this series providing new industry insights, trends and analysis on how AI is transforming organizations. Find them all here.
“We’re experiencing higher-than-normal call volume. For faster service, please use our automated assistant…”
Customers of banks, insurers, brokerages and many other financial services institutions have grown used to hearing such invitations to engage chatbots that can help with questions and transactions. In the coming months, we’ll interact with many more — and not just with rule-based agents with cuter names and better cartoon or stock photo faces.
As part of a surging new embrace of AI, financial services companies are joining others developing next-generation autonomous, cloud-based helpers. Lifelike 3-D avatars that would be at home in a video game combine Large Language Models (LLMs), machine vision, speech and text analytics and other advanced technologies to deliver warmer, more personalized 24 x 7 service — including a growing ability to “read” and respond to customer emotions with appropriate actions and recommendations.
Interest and adoption in AI are booming, finally
More intelligent, empathic and “human” chatbots are the emerging public face of a major shift quietly taking place worldwide in front offices, middle operations and back offices across the financial services industry (FSI). Amidst volatile industry and global conditions, researchers say a diverse array of firms in the sector are adopting AI in record numbers to seize new opportunities. IT and business leaders aim to drive data-driven innovations that will improve customer experience, reduce risk and fraud while creating operational efficiencies and reducing costs.
Figure 2:
The number of financial services firms and executives who view AI as crucial for competitiveness has risen sharply. In the latest “State of AI in Financial Services” global survey by NVIDIA, 64% of the 500 FSI professionals polled agreed “my executive leadership team values and believes in AI,” up from 36% a year ago. Some 58% said “AI is important to my company’s future success,” up from 39% in 2021.
It’s not just talk. Between 2021 and 2025, IDC forecasts that banking firms will invest more in AI than any U.S. industry except retail. The two sectors combined are expected to account for 28% of domestic spending on AI. Indeed, the NVIDIA survey found FSI companies are scaling AI-enabled applications into production faster than ever. One-fifth reported six or more use cases.
AI expenditures in securities and investment firms will see a 30% CAGR through 2025. IDC
From “laggards” to AI champions
Until now, adoption of AI in financial services has trailed other industries. In a recent Accenture assessment of AI maturity, banking and capital ranked last. Insurance fared better, but still below average (See Figure 3). In general, newer fintechs have led adoption, according to Deloitte, with most others limiting activity to small-scale pilots and limited use cases. Big organizations have generally been more active. That’s changing, however, as leaders across the industry recognize AI as the fuel for growth. One indication: a sharp drop in the percentage of FSI respondents viewing themselves as AI “laggards” in the latest “State of AI” survey. (Even so, more than two-thirds believe their company is underinvesting in AI.)
Figure 3:
What’s changed? One major factor is continued pressures to cut costs, grow revenue and make the most of capital during a global economic funk that many in the industry expect won’t lift until 2024. Another big driver: The need to retain and upsell current and new customers — especially the growing, post-pandemic group that prefers to interact via mobile and other frictionless, branchless, people-less ways.
Nearly half of survey respondents say AI will help increase their organization’s annual revenues by at least 10%
More than a third believe AI will also help decrease annual costs by at least 10%
As a result, century-old institutions and new fintech firms — disruptors and disrupted alike — are turning to new cloud-based environments and accelerated processing optimized for AI. The overall goal, leaders say, is to continue innovation, digital transformation and hyper automation efforts that more effectively use data to help firms ride out the current turbulence and position themselves for a competitive future.
The key role of cloud-based infrastructure and accelerated processing
AI, especially real-time workloads, has a notoriously voracious appetite for huge compute power, fast networking and storage for massive amounts of data. As they progress, many financial services institutions are discovering they lack the infrastructure and resources needed to train, run and deploy ML, NLP and other AI-driven applications easily and economically at scale across multiple delivery channels, organizations and continents.
“One of the things we had learned from research is that the larger the model, the more data you have and the longer you can train, the better the accuracy of the model,” explains Nidhi Chappell, Microsoft head of product for Azure high-performance computing and AI. “So, there’s definitely a strong push to get bigger models trained for a longer period of time, which means not only do you need to have the biggest infrastructure, you have to be able to run it reliably for a long period of time.”
That’s where “AI-first” infrastructures come in. The latest advances combine elastic, cloud-based global environments with GPU-accelerated resources and optimized AI/ML algorithms, routines and libraries. Infrastructure as a Service (IaaS) and on-demand acceleration and virtualization can provide high throughput, low latency, scalability, more efficient runtimes and increased model accuracy that can benefit many FSI workloads, especially those running in real time.
Fortunately, many financial firms are well along in efforts to move resource-intensive AI workloads to cloud while keeping sensitive data on premise (Figure 4). Almost half of all AI projects run on hybrid infrastructure (Figure 4). Many are motivated by the desire to shift investments from Capex to Opex, and the realization that a reinvented business demands far deeper involvement in a broader digital ecosystem outside their own organization. The quickening shift to cloud AI makes data portability, MLOps management and software standardization across cloud and on-prem instances a strategic imperative.
Figure 4:
Increased use of public cloud for AI has another attraction: It lets financial services firms take advantage of new offerings that can further accelerate cost-effective production AI.
For example, Microsoft and NVIDIA last fall announced a massive, cloud-based supercomputer to help global enterprises with AI development, deployment and scaling of AI including large, state-of-the-art models like LLMs. The system combines Microsoft Azure’s advanced supercomputing infrastructure with tens of thousands of NVIDIA A100 and H100 GPUs, NVIDIA Quantum-2 400Gb/s InfiniBand and full stack of NVIDIA AI Enterprise Software to maximize application performance and accelerate the path to production AI deployments. In addition, Microsoft previewed two more new purpose-built-services: massively scalable virtual machines to accelerate generative AI, and a pay-as-you-go, managed parallel file system to speed AI and HPC workloads.
Another important development: industry-and sector-specific clouds, accelerators and business services for FSI. Advocated by Forrester and other top consultants, this emerging group can be easily deployed and customized, further speeding time-to-value. And there’s another big benefit, notes Deloitte. The approach frees businesses to focus precious development and data science resources on critical areas of strategy and business differentiation. At a time when the biggest AI obstacle for FSI (and others) is finding skilled workers, that’s a huge advantage.
All these cloud-based services for AI bring another huge benefit: They let enterprises access the latest technology, on demand. With technology capabilities rapidly doubling, the ability to extract immediate value is a huge competitive advantage. In contrast, on-prem technology investments typically stay in place for 4-5 years.
Top use case: NLP-driven customer experience
FSI firms are leveraging AI-optimized cloud environments to deliver top-line and bottom-line results more quickly. Dozens of major use cases (and hundreds more specialized ones) show how accelerated, purpose-built infrastructures and high-performance computing (HPC) can help financial services respond to increasingly demanding customers and sophisticated bad actors.
Unsurprisingly, popular use cases include AI risk management and compliance, fraud prevention and detection for insurance, identity and anti-money laundering (AML). Pricing is another high-value application; for example, startup Riskfuel is pioneering the use of deep neural networks to learn complex pricing functions used to value OTC derivatives. Yet the top priority for many firms, within and beyond FSI, is improving and evolving customer experience.
For example, NLP capabilities added to live agent transactions can gauge caller sentiment and file urgency, then respond by routing a call to the appropriate department more quickly and accurately. Or provide language translation to solve language barriers between a caller and agent.
And by combing through massive data sets, NLP can conduct sentiment analysis to identify complaints, reviews and mentions across multiple touchpoints, spot and understand patterns in customer behavior undetected by humans — even learn and adapt to site visitors by saving preferences or tailored custom content. To further humanize interactions like a human operator, some digital agents may even tell jokes like other digital helpers: Example: “Where do fish keep their money? In the riverbank…”
Meet Violet
One of the most exciting innovations emerging is AI-powered virtual assistants (avatars). Interactive, self-learning agents with more expressive photo-realistic or animated faces can provide intelligent, friendly interactions on mobile phones, computers, kiosks, and virtual environments to attract and retain financial services customers through personalization and recommendations.
Like their familiar rule-based ancestors, the new breed of service agent can speed up many routine transactions and queries. They can quickly understand, for instance, if a chat session or voice caller wants information such as a balance update or payment confirmation.
But the new real-time, cloud-based avatars go way beyond existing chatbots. Because they can leverage LLMs, NLP and ML, the questions and tasks they can manage aren’t limited to pre-written scripts. And increasingly, these 3D avatars can use non-verbal cues like facial expression and eye contact to enhance communication and understanding of your requests and intent. Some are working to exhibit empathy – an emerging holy grail in banking and other sectors.
Consider a new interactive avatar named Violet. Deloitte used the NVIDIA Omniverse Avatar Cloud Engine (ACE), a collection of cloud-native AI microservices and NVIDIA Tokkio, a domain-specific AI reference application for developing fully autonomous customer service avatars, to create three versions: Customer Service, Intelligent, and the human-like UltraViolet, pictured above. Demos at industry events have been a big hit. More financial services firms are expected to bank on such digital helpers to deliver a more natural, comfortable and engaging — and profitable — customer experience.
“A special offer just for you…”
A popular cornerstone of improved customer experience is a new generation of AI-driven recommenders. More financial services firms are implementing these systems, which use predictive analysis on stores of customer data (payments, balances, credit history and score) to deliver personalized insights, intelligent credit scoring, proactive customer support and new services and offers to individual customers.
Developing and training ML recommenders is time-consuming and complex. It requires staff with specialized skills such as data science. Moving from legacy AI infrastructure and CPU-based processing to a cloud-based, with pre-built, pre-tested models delivers the huge computational resources needed for faster processing and training, helping to accelerate and increase ROI.
Bottom line
As an industry, financial services have been slow to bring AI into its mainstream. But the long history and significant resources that popularized the commercial use of computers, ATMs, mobile banking and algorithmic trading seems sure to develop AI innovations that will help other industries reimagine their own customer experiences — and entire businesses.
Microsoft and NVIDIA have a long history of partnering to bring financial services s firms leading-edge technology to support AI and ML, Microsoft Azure purpose-built AI infrastructure and the NVIDIA AI platform provide scalable, accelerated resources needed to run AI/ML algorithms, routines, and libraries.
Go deeper
Azure AI Infrastructure
Reinvent the financial services experience | Azure
AI Solutions for Finance Industries | NVIDIA
Enhancing Customer Experience Using Natural Language Processing
Building Cloud-Native, AI-Powered Avatars
VB Lab Insights content is created in collaboration with a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.
[ad_2]
