AI dividends: Financial firms want to capitalize on gen AI. Here are the best places to start.

Like pretty much every industry, banking, insurance, and other financial sectors are eagerly exploring, if not outright embracing, generative AI. For even the most regulated fields, the productivity, creativity, and innovation promised by gen AI is too great to ignore.

Just ask the 350 banking executives surveyed last year by The Harris Poll and Google Cloud. Among them, 96 percent said that senior leadership was actively involved in gen-AI decision making and strategy. And just over four in five said they were already in the proof-of-concept stage or actively piloting use cases, while nearly all respondents (99%) shared they are actively hiring AI roles. Among the most popular projects nearly half (49%) are seeking to tap into gen AI to drive cost savings and efficiency (49%).

With all this interest, the many financial services leaders we regularly meet with are still searching for the best ways to guide their generative AI investments and upskill their teams. Some 47% of banking executives say they are in the proof-of-concept stage and 35% are testing use cases, but from there, the path forward becomes less certain. Experimenting with pilots is relatively easy compared to the harder reality of implementing and scaling gen AI capabilities to capture tangible business value.

Each of us have spent decades leading product and transformation efforts at some of the biggest banks in the U.S. before joining Google Cloud, so we know just how challenging it can be to roll out new technology at speed and scale. Having heard in our surveys some of the chief concerns financial institutions face when it comes to gen AI — specifically around governance and security, talent, and customer experience — we wanted to share some thoughts on where to focus to get the best results now.

Investors and institutions are always looking for alpha, that little bit of edge that can compound massively over time. For the finance industry, generative AI represents one of the greatest alpha opportunities in a generation — and not only for your business’s returns but also for the efficiency of operations and the delighting of customers.

Now is the time to make those serious AI investments, and here’s where to start to see the best returns.

1. Gen AI journeys in banking start with governance

When it comes to implementing gen AI, a critical and often overlooked starting point to unleash innovation is governance. While banks are highly motivated to incorporate gen AI technology to help them transform their business, they often run into unexpected roadblocks when it comes time to get these initiatives approved.

In particular, one of the biggest hurdles that doesn’t get discussed during planning for technology projects is the need to develop governance structures for navigating security controls and risk assessments. It’s a classic chicken or egg dilemma — you can’t use AI to improve business operations and processes until it’s approved and passes through existing governance processes.

Unfortunately, governance is one of the most common reasons that technology initiatives never move beyond experimentation. It's a necessary part of scaling any solutions and services — gen AI or not —and bringing them to production. Again and again, we’ve seen financial institutions waste countless hours and resources exploring exciting and strong technology use cases, only to have them stopped or abandoned when it reaches the point of gaining approvals.

Ultimately, no experiment will ever be effective if it can’t enter the next phase of adoption. Transforming governance, therefore, should be addressed concurrently as organizations seek to identify the best use cases for gen AI.

Many existing governance models were built up around legacy waterfall software development models. As a result, governance committees are used to evaluating technology projects that follow strict, sequenced lifecycle processes, with discussions sometimes taking place years in advance. Whereas, the future of software development, especially in the AI space, will need to be agile and flexible to keep pace with changing requirements and customer expectations.

It’s important for banking organizations to take the time to reimagine governance structures and redefine their processes to enable alignment, accuracy, and accelerate approvals. Banks should evaluate their current governance models and frameworks to consider what controls will be needed to support the rapid adoption and scaling of new gen AI capabilities.

2. Your first AI hires aren’t who you think

Another important factor to maximize the impact of gen AI technologies and tools is having the right people to make it happen — and the roles to hire first might not be what organizations expect. The rush to acquire new AI expertise and technology skills aren’t necessarily the right fit for banks that face a challenging regulatory landscape and intense scrutiny to deliver AI responsibly and securely.

While all organizations should be prioritizing responsible AI, the banking sector is particularly concerned about ensuring that gen AI is accurate and secure. Highly-regulated industries, such as financial services and healthcare, have more regulatory oversight and a need to comply with stringent data and privacy regulations. Technology missteps and errors can carry serious consequences for financial institutions, placing even greater pressure on them to fully understand and mitigate the potential risks of gen AI.

The first, second, and third lines of defense are already overwhelmed with the sheer amount of risk to manage, particularly in large global financial institutions. For a sense of scale, one of our customers shared with us that they have almost 10,000 controls — just within North America — and the vast majority of them are still manually controlled and monitored.

Too often, organizations tend to focus their attention solely on hiring AI and technology experts and forget to consider the other specialized skills that contribute throughout the entire development lifecycle. Roles in these areas will also need to be increased, and quickly, to ensure gen AI projects can move into production quickly and smoothly.

While hiring new AI talent is undoubtedly a crucial factor for achieving gen AI success, financial institutions will also need to focus on investing in new talent to support governance, particularly “second line of defense” roles in compliance, legal, operational and fraud risk, and the office of the CISO.

In fact, it’s already happening at a lot of banks. Our survey found that the top roles executives said they were hiring for were in “AI quality assurance and testers,” with 34% saying their firms were hiring for such positions. Strategy consultants (32%), AI product managers (30%), cybersecurity experts (29%), and DevOps engineers for AI (29%) were the next most popular roles.

The top skill, after “knowledge of the cloud,” that the executives were seeking was “understanding of AI ethics, regulations, and compliances,” which almost half (49%) of firms were seeking.

Hiring isn’t the only answer, though. Organizations should also take steps to retrain and educate all talent, new or existing, about AI and cloud technologies. In risk-averse businesses like banking, a clear understanding of how technology works and the risks associated with it can be the difference between driving change or stalling out.

Business executives are typically the gatekeepers for innovation. If they don’t fully understand the technology or the risks associated with it, the assumption is that it’s high risk. It can be short-sighted to believe that talented engineering teams and enthusiasm for gen AI will be the main engine driving new innovation. Without the support of all your governance partners to keep projects moving forward, most organizations will be unable to gain the anticipated value from their initiatives and may find the final value falls short of expectations.

3. Start with easy wins to gain long-term momentum

As banks move farther along and start shaping their new gen AI strategies, they will need to have a strong vision and strategic roadmap to follow, including a clear view of goals and a realistic assessment of how long it will take to enable capabilities, including data and technology requirements, hiring and training talent, and the operating model.

For any institution looking to get started with gen AI it’s best to focus on short-term wins that require medium to low effort before trying to tackle large and more complex use cases. A unique benefit of gen AI’s ability to drive efficiencies and cost savings is that banking organizations can target areas with the highest costs and burden first.

For instance, services is an area where we’ve seen all major banks consistently increasing headcount over the last decade. Front-office operations in branches and contact call centers, have long been among the slowest areas to transform in banking. The experience for customers who need support from in-branch personnel or human service agents is not much different than it was 20 years ago, with long lines and wait times to get answers, help, and information as bank workers search across systems.

These call centers and front-office activities represent an enormous burden on annual spending. Each second spent to handle a call, for example, can add up across the firm to as much as $2 million on an annual basis. Depending on the size of the institution, reducing the average call by even a single second could realize millions in savings over a course of a year.

It’s unsurprising then that we see many financial institutions starting with internal chatbots that can help agents find information and address issues faster, or pursuing other forms of gen AI-powered enterprise search capabilities to help workers across the organization get answers faster to jumpstart their work.

This approach comes with several advantages. Gen AI assistants are useful for helping to take on some of the more repetitive tasks that take up the most time and freeing up time for people to focus on revenue-generating activities. At the same time, tackling areas that monopolize time and resources upfront can deliver funding for future innovations, enabling organizations to reinvest savings into other key areas.

As an added benefit, we also find that internal-facing use cases like AI-powered enterprise search are easier and faster to get through approvals as they tend not to involve sensitive customer information. The risk of applying gen AI to data that might include sensitive personal information like a social security number is exponentially higher than accessing documentation that outlines an internal procedure on how to open up a checking account.

A good rule of thumb to keep in mind as you are building your roadmap is to start with use cases that support employee processes first. This allows organizations to get comfortable with the basics of gen AI and find their footing with dealing with governance and responsible AI before trying to apply it in more complex areas and personalization opportunities, such as business intelligence analytics or external customer-facing activities like marketing.

With these foundations in place, banks, insurers, and financial institutions will have the ability to take that success and use it to explore even bigger opportunities down the road. The banking industry is on the precipice of an entirely new arena, where organizations will be expected to have a distinct understanding of who their customers are and differentiate themselves through their services and experiences — and do it for a lot less.

All together, these insights can help to create a springboard for future innovation, setting up banks and financial services institutions in a position to move faster, realize value, and continue bringing new opportunities to build a competitive advantage.