As the race for technology and AI innovation continues to shroud the financial services industry, two industry players call for patience
How the financial services industry will develop its technology and automation processes has long been the centre of many conference panel discussions, articles, and reports. The industry is gripped with innovative fervour, but maybe it is counting its chickens before they hatch.
Demetry Zilberg, chief information officer at Alter Domus, is keen to call for patience. He explains that, “the hype cycle for AI is quite intense and it is still early days.”
For Zilberg, getting technology correct is more important than innovating quickly. “You have to get the technology approach right, otherwise we have a very bad stability outcome,” he says. “There are a number of stories about large financials being compromised from a cyber perspective, because they messed up their implementation onto web services. It’s better to be some percentage slower, than to have a bad outcome in a production live environment.”
The hyperfocus on developing technology has stressed the fibres of the industry, Bill Conner, CEO of Jitterbit believes. He says: “Businesses are already grappling with complex IT landscapes, managing an average of nearly 1,000 applications with only a 28 per cent integration rate. The integration demands of AI systems further strain these overstretched teams. These challenges fall into three main areas: ensuring high-quality data, overcoming technical hurdles, and navigating significant operational changes.”
On the grill
“The financial services industry has a wealth of information and holds immense potential for AI-powered advancements,” Jitterbit’s Conner begins. “But data quality, security, and integration pose significant roadblocks and related risks to initiate AI processes.”
Ensuring firms have data that matches quality just as much as quantity, is central to ensuring technology can work for firms and their clients. Conner continues to explain that, “Ensuring clean and compliant information across disparate sources is crucial for accurate analysis, making data management a critical first step in unlocking the true power of AI in banking. This is a large undertaking that should be measured and paced with the company’s best interests in mind.”
Speaking to Alter Domus’ Zilberg on the Fourth of July, he takes great pleasure in comparing his day’s celebrations with technology. “I’m going to be grilling later. Data is like the ingredients for grilling. I went and bought a good brisket, I bought some pretty fancy pork ribs, and I hope I’ll have a good outcome in a couple hours when it comes off my grill. It’s the same with AI implementations, if your data is garbage, garbage is going to come out.”
Turning his mind back to the industry and away from his grilled meats, Zilberg remains confident in his company’s own ability to deal with data. He explains: “Data can be tuned to the specific use case of the output of your AI application or tool, but what is critical is data privacy and compliance. But I don’t think it’s a challenge for us, it’s a strength.”
Tech’s got talent
So, what are the challenges?
“Finding talent is an all out competition,” Zilberg exasperates.
“Finding people capable in the space is remarkably competitive and the salaries and the incentives are just insane, unlike anything I’ve seen in my career.”
The lack of talent is not enough to keep pace with the accelerating levels of innovation, Zilberg explains: “There’s not enough talent. There are more ideas in the backlog than we can implement, probably by a factor of 10. It goes back to how do you build teams, how much do you do in house, who do you partner with?”
With the biggest companies vying over the best and brightest talent in a limited pool, companies have to find ways to combat the issue of manpower should they not be able to attract and retain key staff.
Zilberg points to partnerships as one solution to escaping the challenges that come with the dearth of talent.
With smaller teams under greater pressure to deliver, and deliver quickly, mistakes can happen — with some of them could be fatal for companies.
Zilberg warns: “You cannot screw up on data privacy because the reputational loss is irreversible. It goes back to going a little slower and being deterministic, quite careful, and testing before releasing into production.”
Jitterbit’s Conner is slightly less pessimistic and argues that “by taking an AI-infused approach, financial institutions can safely and proactively adopt enterprise AI capabilities without jeopardising business operations or disrupting current financial systems.
“This delivers customer choice and can help accelerate AI adoption without forcing a one-size-fits-all approach. This is even more important in the highly sensitive financial space.”
Conner remains wary however and urges “every financial services company [to] deploy AI at its own pace, looking at the ways in which it can be utilised within the needs and constraints of their specific environment.”
The underlying message, however, remains as Conner puts it: “AI is about empowering an evolution, not a revolution.”
There should be care, not clamour.
