Financial services can bank on the potential of generative AI and automation

The financial services industry has a golden opportunity to transform itself with generative AI (GenAI). The McKinsey Global Institute estimates an annual revenue increase of $200-$340 billion (equivalent to 9–15% of operating profits) for banks that leverage the technology.

However, financial services organizations must surmount key challenges to fully capture the technology's potential. Data security remains an anxiety for banking leaders. They aren't sure if they can trust GenAI to handle their most important client interactions, especially under tight regulatory scrutiny. And many are also grappling with finding and prioritizing use cases that will drive rapid, proven, and transformative value—and making fast progress on capturing this value.

By using automation to release the power of GenAI, first movers are successfully meeting these challenges—generating transformative results in many areas of their business, from enhanced customer experiences to streamlined workflows and beyond.

In today’s blog post, we’ll identify key areas where GenAI and automation have already made a real impact in leading banking organizations and provide effective roadmaps and strategies for doing so in your organization.

Where to focus first: profitable use cases for GenAI in banking

GenAI has so much across-the-board potential to transform the financial services industry, that first movers in GenAI, enabled by automation, have already yielded results in a range of core areas:

Regulatory compliance

Regulation is a big cost center for banks, with a third of industry executives spending 5% of their annual budgets on compliance alone. Fortunately, GenAI is improving outcomes and lowering costs. Financial services firms are already using the technology in automated processes that continuously monitor transactions for potential fraud and to generate suspicious activity reports. GenAI is also starting to be used as an assistant for compliance professionals, comparing policies and answering questions about key regulations.

Audit and controls

GenAI is transforming internal audit and control frameworks from defensive to proactive. It’s capable of analyzing massive data sets, improving testing accuracy, and increasing testing frequency. Using automation to streamline data gathering and preparation, Turkish bank Yapi Kredi has made GenAI a core part of its auditing process, using it to verify signatures and make other checks.

Risk

GenAI has diverse use cases across credit, cyber, and climate risk. It can inform lending decisions by summarizing customer data, automate the drafting of credit memos and contracts, and generate default and loss probability estimates. GenAI can also create high-resolution risk visualization maps and automate data collection for transition risk assessments.

Developer productivity

Helping to create new code from scratch, GenAI significantly reduces a developer's manual workload. It can be asked to review code, quickly spot issues, and even suggest how to fix them. At banks like Goldman Sachs, GenAI is making development faster and more cost-effective, supporting developers and significantly boosting their productivity.

Thinking bigger

While existing GenAI use cases can create real value, they’re only a fraction of what the technology is truly capable of.

Banking leaders need to think in the long term. Stanford research shows that large language models (LLMs) powering GenAI are increasingly able to simulate realistic human behavior and interactions. The progress made in the last year has been extraordinary and the future potential is vast. GenAI could augment almost all banking functions, from customer service to compliance.

In the next few years, GenAI will redefine the customer relationship in banking. Customer service agents using AI can supercharge their productivity as they handle multiple chats at once and spend less time per chat. GenAI will personalize customer experiences while simultaneously assessing credit risk, flagging potential fraud, and handling high-volume processes.

But even these predictions are limited by today’s thinking. The truth is that tomorrow’s most valuable GenAI banking use cases haven’t even been imagined yet. One of the industry’s most valuable resources has been the scale of its unstructured data assets. Due to costs and manual constraints, these are mostly untapped. But GenAI holds the potential to finally unlock them, creating countless new applications and possibilities.

However, before banks can run, they need to learn how to walk with GenAI. Now is the time to strategize and lay the best foundations for GenAI success.

From idea to implementation: operationalizing GenAI

Rather than just chasing a trend, banks need to strategically identify key areas where GenAI can make a genuine impact and move the needle on their bottom line. As always, deployment should be based on three criteria: technical feasibility, business case value potential, and risk. For any AI use case, you should ask yourself:

• Can GenAI complete the task safely, consistently, and reliably?

• What is the value, and does it justify the considerable investment GenAI demands?

• What are the consequences if something goes wrong?

Every use case is different, and every bank will have its own unique risk appetite, and responsibilities to clients, regulators, and shareholders. However, the success of any use case hinges on the foundations (technical and cultural) put in place by the bank.

Before embarking on any GenAI deployment, you should do the following:

1. Build your foundation in the cloud

GenAI is predominantly a cloud technology. It requires substantial compute resources for model training and execution, including large volumes of data and significant computational power. What's more, banks often need to analyze large-scale, continuous streams of data in real time for tasks like fraud detection or market predictions. Traditional infrastructure will struggle to meet these demands, but the scalable and distributed nature of cloud technology allows it to handle such tasks easily.

Banks that haven’t laid the foundation for cloud and cloud services will, in turn, hit a major roadblock when trying to deploy and scale GenAI. The direction of travel is clear—the latest GenAI advances and updates will only be available via the cloud. That’s why cloud adoption across your tech stack is so important right now. Even leaders in cloud adoption will need to significantly scale up their cloud resources to prepare for the GenAI workloads of the future. As AI technology advances and proliferates at increasing speeds, choosing a rapidly scalable cloud service model, a reliable cloud provider, and cloud-based automation platform are critical.

2. Realize GenAI potential with automation power

Automation is the chief enabler of GenAI.

GenAI is like a brain without a body. It can think, reason, and even talk to you. But it can’t do anything by itself. And it has an in-built air gap with your most important enterprise systems— a gap that can only be filled with APIs, connectors, and automation tooling. GenAI’s true value is realized when it’s integrated, not just with your human employees, but with automation technology as well.

Automation serves as the intersection between traditional platform technology (on which banks are built) and AI. It transforms the value potential of GenAI into real business value, enabling impactful use cases and greatly accelerating time to value with the technology. Working in concert with automation, GenAI can finally take action, update systems and records, and respond rapidly to customer requests.

Cloud-based automation platforms that incorporate GenAI open the door to powerful features and capabilities. Fundamental weaknesses of GenAI technology are mitigated or even resolved.

One example is the challenge of business context. GenAI models are typically generalist in nature, being trained on huge general-purpose datasets. This helps GenAI perform well across many different tasks, but accuracy suffers in highly specific scenarios. Banks can spend a long time educating their GenAI models with relevant business context. Or they can use automation to quickly gather the relevant business data needed to ground the model in their unique business context.

Security and reliability are also enhanced. GenAI models are often described as ‘black boxes’. They ingest data to help generate accurate responses, but they don’t tell you what specific data they used or how. With so little transparency, it’s understandable that banking leaders might shy away from GenAI.

Fortunately, the UiPath AI Trust Layer provides a solution. It gives banking leaders clarity into how their GenAI models are being used and shows what data is impacting their decisions. The UiPath AI Trust Layer also provides access controls, helping you decide exactly who in your organization can leverage GenAI models while monitoring their usage.

3. Redefine productivity for your people

Despite the proliferation of AI, the human factor remains critical. Banking leaders need to take a people-first approach rather than a technology-focused one. AI is, first and foremost, a tool to enhance human labor and productivity rather than replace it.

That’s why humans need to remain a core part of AI-augmented processes. Always leverage humans in the loop to review the most important AI outputs and decisions before they’re carried out. This will help increase the scope of GenAI use cases while maintaining the same or even greater compliance and oversight. You can improve efficiency further by getting GenAI to check its own work, including through source citations and risk scores, to cut the reviewer's workload.

Consider also how GenAI and automation can level up employee productivity and expand their capabilities. UiPath Autopilot™ is a great example of how this can be done. Through simple natural language prompts, employees can trigger powerful automated actions through Autopilot, which blends specialized and generative AI to understand what employees want to achieve and how it can help them.

You can also look at investment banking leader Lazard, which managed to automate the time-consuming process of creating pitchbook reports for its bankers through GenAI and UiPath automation. Not only did this free up employees to focus on cultivating client relationships, it helped them make faster and better business decisions thanks to GenAI summarization.

As mentioned previously, the most valuable GenAI use cases are yet to be discovered. The only way that will happen is if your people can capture the art of the possible—understanding the technology, its value, limitations, and how to turn its potential into action.

AI education should be a top development goal for everyone—from agents to the bank’s top executives. To make those disruptive, game-changing discoveries with AI, you first have to augment your organizational imagination when developing AI use cases.

Banks must build today for tomorrow’s AI

Organizations can't afford to wait and see what happens next. We’re past the turning point and have entered a new business paradigm with GenAI. Banks dragging their feet over implementation are leaving incredible value on the table and fighting the gravitational pull of the future.

As the speed of AI development is so rapid, there's a need for banks to learn and mature alongside the technology or risk being left behind. It’s imperative, then, that banking leaders invest now to understand the technology, upskill their people, and build up their automation and cloud capabilities.

However, banking leaders should feel bullish. The financial services industry has long been ahead of the curb, setting the standard for adopting and transforming themselves using cutting-edge technology. Banks are ideally placed to be our AI pioneers.

For more advice and practical strategies to leverage the power of GenAI in your banking organization, download our eight steps guide to operationalizing AI.