The financial services industry has evolved rapidly since the global financial crisis of 2008 spawned the rise of fintechs and, with them, a customer base with increasingly demanding digital expectations. The pandemic has accelerated digital behaviours even further, seeing technology leap over brand trust as the key battleground on which new customers are won.
Specifically, artificial intelligence (AI) is becoming the key differentiator. With digital having thoroughly eclipsed physical branches as the dominant medium of engagement for financial services, companies must now focus on personalising their customer experience through the likes of chatbots, virtual assistants and video banking. If they are not able to provide these kinds of AI-enabled experiences, their customer service will become inferior.
“If you want to drive customer lifetime value and retention, cross-sell products and deliver better customer service, it has to be in a digital environment and ecosystem, and to be best in class, it must be AI-infused,” says Kevin Levitt, industry business development lead for financial services at NVIDIA. “Financial services firms undoubtedly now see the importance of AI to their competitive position, and the need, therefore, to set an enterprise AI strategy.”
Thanks to the capabilities of companies like NVIDIA and Hewlett Packard Enterprise (HPE), the computing power required to deploy AI in enterprise environments has caught up with the theory, accelerating AI projects. While some are still only discussing AI, and at the other end of the scale it is almost pervasive in some startups, most companies have reached a stage of having now enjoyed success with several smaller AI deployments and wish to scale those quick wins into something larger.
There are, naturally, technology challenges involved with scaling up AI in an enterprise, not least establishing a coherent data strategy to underpin deployments, but the greatest challenges are organisational. A trifecta of personas is key to the success of any AI deployment – IT, data scientists and line of business – but contrasting roles and objectives can often mean they are not only speaking different languages but that there is active tension.
The efficiency of the data scientists, a highly costly resource, is critical to the success of deployments. If the data science group cannot work effectively between IT and the line-of-business project owners, the project is likely to flounder. They require a strong operational path that enables them to use the tools they want, how they want, and access the resources they need, without usual IT blockers like raising a ticket or waiting for access.
“Collaboration is vital,” says Adrian Lovell, chief technology officer, financial services industry, at HPE. “If you’re not on top of that sort of dynamic, in the best case it’s going to take you significantly longer to derive any value from your efforts, and in the worst case you’re not going to get anything out of the lab and it’s just going to die there as a science project. You’ve absolutely got to find a way to break through the friction and facilitate the collaboration that is required.
“It’s about understanding that nobody wants to cause a problem. The IT people may simply feel protective over what they do for a living and don’t like the idea that somebody else may want to try and change that. The data scientists may have come from academia or another industry where they were able to do everything they always wanted to do and are now adjusting to a structured, corporate environment without as many of those freedoms. And then there are the product owners saying, ‘I’m trying to make money here, just get it done.’”
Amidst these organisational tensions there is also the creeping issue of shadow AI, whereby the pressure on lines of business to leverage AI-enabled applications has resulted in siloed, fragmented implementations, some in the cloud and others on-premise. There is no AI infrastructure at scale across the enterprise, leaving financial services companies struggling to get to market with the AI-enabled services that will help differentiate them.
“All these challenges are bubbling to the surface,” Levitt adds. “That’s the confluence of what happens when these three important pieces of the puzzle are not organised and talking to each other. Ultimately it surfaces a broken AI enterprise because there is no strategy to go with it. That’s where I would say there’s a lot of tension but also opportunity in terms of the conversations that we’re having across the executive suite today within financial services.”
Financial services organisations are looking to partner with leaders in this space to help them understand how to deploy effective AI infrastructure at scale to get the most value out of one of their most valued resources: their data scientists. Many are turning to a partnership between HPE and NVIDIA, which together can provide an enterprise AI infrastructure at scale that’s full stack, from the hardware up to the application suite, enabling companies to operationalise AI-enabled applications for any variety of use cases.
Waiting for a model to train for days or potentially a week is an incredible waste of a valuable resource in a data science team. The combined HPE and NVIDIA solutions help financial services firms wherever they are in their AI maturity, from those already seeing the benefits and wanting to scale that capability, to the laggards which are now realising they can’t compete in the marketplace unless they develop a formidable AI enterprise strategy. Procuring solutions as-a-service can also assist the challenges between the trifecta, with HPE providing all solutions as a service through HPE Greenlake.
“It’s all about the partnership when you are tackling these sorts of challenging problems,” says Lovell. “I don’t think anybody would attempt to try and do this by themselves. We like to work with our friends at NVIDIA because we believe in bringing what the best of breed is to our financial services customers to ultimately try and address those fundamental problems.
“We’ve been working with NVIDIA to help the trifecta of personas involved in AI deployments understand each other’s point of view, because that’s really not a technology problem. Technology can help the problem but fundamentally it’s about understanding each other’s point of view and seeing, frankly, that everybody needs to work together or nothing will be delivered. That’s all that matters to the business at the end of the day: getting it delivered.”
For more information, visit hpe.com/uk/getahead