Delivering a premium customer experience has always been essential to financial services. But in today’s rapidly changing environment, financial institutions need to do more than maintain the experiences they’ve developed.
As customers increasingly seek out personalization, generative AI and large language models (LLMs) can help cutting-edge fintech startups and incumbent banks alike provide the instant answers and solutions that meet growing customer demand. The industry has already caught onto the potential of these technologies: Nearly half of the financial institutions and fintechs surveyed in NVIDIA’s State of AI in Financial Services report say they’re already using AI to improve their customer experience.
By building on the right platforms, like the AWS AI/ML platform powered by NVIDIA’s accelerated computing platform, both fintechs and incumbents can leverage generative AI and LLMs to enhance the customer experience in fundamental ways — from sophisticated chatbots and digital avatars, to hyper-personalized content across marketing and product education, to automated text summary and document search that boost efficiency and prevents fraud.
Meeting customer needs with chatbots and digital avatars
Traditionally, self-service options like chatbots or virtual assistants have been confined to a pre-written script, a stifling limitation when dealing with user’s real-world issues. “Ultimately, if you ask enough questions, the predefined scripts will end, and you’ll have to be transferred over to a customer service agent,” says Kevin Levitt, Head of Global Industry Business Development for the Financial Services industry at NVIDIA.
But thanks to generative AI and LLMs, chatbots are becoming increasingly sophisticated and easy-to-use — so much so, in fact, that they may even be preferable to their human counterparts, especially when discussing sensitive matters. Today, companies can use tools like Cleo, a chatbot that dispenses advice for personal wealth management. By leveraging Amazon SageMaker and sitting within the AWS AI/ML platform powered by NVIDIA’s accelerated computing platform, Cleo can speak to company policies, product requirements, and inquiries regarding a customer’s personal situation — all within a familiar and comforting conversational tone that’s equipped with the appropriate conversational guardrails.
Beyond chatbots, virtual assistants and digital avatars like the ones brought to life by NVIDIA Omniverse Avatar Cloud Engine (ACE) can understand user needs based on their audio input and make intelligent recommendations in real-time in order to provide a more empathetic experience. In financial services, building this kind of intimate connection with customers is especially important.
By using digital tools to handle the bread-and-butter customer service requests, companies can then reserve their customer service representatives for truly unique cases. In those situations, too, AI is here to help. Traditionally, a large part of an agent’s responsibility during a customer service call has been taking comprehensive notes. Now, they can leverage speech-to-text AI to produce a transcription that doesn’t just document what’s happened, but also powers other AI features, like recommendation systems, for both customers and the agents to understand how to reach the most satisfactory resolutions.
Empowering engagement with hyper-personalized content
It’s well-known that AI-powered applications can recommend personalized ads to customers online. In addition, these same algorithms can go one step further and determine which impressions are not just most likely to motivate a conversion — but also most likely to help empower customers to make more informed financial decisions. In fact, by building with the NVIDIA Merlin framework available on AWS, Capital One unlocked a superior attribution method that outperformed alternative models by a wide margin, allowing the company to serve personalized marketing and educational content that went above and beyond typical ads.
To better engage its customers, Capital One analyzed their relevant data and then served them content that best matched their financial journey, whether that was doubling down on saving for retirement, budgeting to buy a home, or creating a more effective investing strategy. This bespoke approach is essential to forging connections with your customer base, says Levitt.
“The generative AI can create personalized advertising copy, personalized emails, and personalized newsletters,” he says. “Obviously, that helps with click-to-conversion rates when it comes to new customer acquisition or cross sales, but it’s also really important in terms of building customer loyalty and retention. With that personalized experience, the customers feel heard and understood.”
Slashing wait times and boosting compliance efficiency
Most home mortgages take up to 30 days to close, and even that’s sometimes a feat. Banks have to analyze and process a mountain of documents, and when key information is missing, they have to communicate with the lenders, who may not fully understand what’s needed from them. In these situations, too, generative AI could take over the rote functions. Doing so could shorten the process from 30 days to a single week.
“It’s not just a boon to the customer or the individual that is looking to purchase the home, but it helps the bank in terms of the efficiency and productivity of their workforce,” Levitt says. “It also means that their human resources can focus on higher-order, more complex processes to create more value for the bank.”
Those same AI-powered tools are also invaluable for in-house processes like fraud detection and prevention, says Sam Edge, Global Head of Fintech for Startups at AWS.
“Generative AI will greatly improve the efficiency of how tech companies and compliance officers perform their duties,” he says. “It’s already helping them scan a document, read the text on that document, and then figure out how to categorize it based on the compliance models that the company may have.”
Despite the promise and potential of this technology, banks and fintech companies must tread carefully to protect their customers and maintain trust. Instead of deploying a model built on the entirety of the internet, Levitt recommends they fine-tune it based on their specific business and its unique market.
And when it comes to that data, companies need to both own the IP and allow customers to opt out of incorporating their data into the training model. Security and confidentiality are the bedrock of financial services, and that’s just as true when it comes to generative AI.
Thanks to the AWS AI/ML platform, powered by NVIDIA’s accelerated computing platform, these technologies are ready to be used by banks and fintech enterprises. The end result is an AI-powered customer service flow that’s better for customers and companies alike.
CTA: Discover how the AWS Generative AI Accelerator is helping startups innovate and scale in just 10 weeks.
