Leaders and the workforce at all levels are increasingly captivated by the immense potential of generative artificial intelligence. In the banking industry, its ability to revolutionize fraud detection, loan approvals, customer service and many other functions is highly promising. However, a hasty and unstructured approach to its implementation can lead to a series of “random acts of digital” that can increase costs and risks and undermine a bank’s potential success with generative AI.
Random acts of digital refers to the sporadic, impulsive, or unaligned planning or implementation of digital technologies. This includes generative AI strategies, solutions or technology without a clear plan and approach to align with the organization’s broader goals, strategic vision, and operational, talent and adoption needs.
For example, a key point of concern in banking with generative AI (or GenAI as it’s sometimes called) is the potential risk in how data is used. Banks, of course, are particularly sensitive to this due to the strict regulatory environment, importance of customer data confidentiality, and reputational risk. A rise in random acts of digital could increase the risk of data breach, reputational damage, or regulatory noncompliance.
On the flip side, not moving forward quickly enough may cause banks to fall behind and miss out on the impressive value generative AI offers. In our recent conversations with different banks, we find they are going down one of two paths: one being so cautious nothing substantive is happening, and the other moving forward as quickly as possible while being very careful as they do it.
- Notably Quotable: What Banking Leaders Think of Generative AI
- Why You Need to Be Experimenting with AI Tools Like ChatGPT
GenAI Perils: Avoiding Unintended Consequences and ‘Use Case Fever’
Generative artificial intelligence can bring considerable change to banking processes, altering everything from risk management protocols to customer interaction patterns. It can facilitate improved credit scoring algorithms, optimize portfolio management, and augment financial advisory services.
However, these benefits can be overshadowed by the fallout from sporadic and uneven planning and implementation.
Addressing the impact of generative AI on banking processes is crucial. For instance, a change in credit scoring due to the implementation of a GenAI solution would require alterations in risk management strategies, employee training, and customer interaction protocols. In banks, where strict governance is mandatory, such changes need to be carefully planned and orchestrated. But banks cannot be so cautious that nothing moves forward.
It’s important to remember that with generative AI, a new “use case” or changes to an existing process will invariably cause the need for change in other areas. Think “interconnected system,” not “silos.”
How 'Use Case Fever' Stalls Progress:
Employees in many companies are so eager to apply generative AI they propose a large number of use cases. But they often fail to win support because they haven't worked through all the risks. This type of frenzy depletes valuable resources.
A well-known, large-language-model, generative AI company shared with us one of the biggest challenges they are seeing in companies trying to create GenAI solutions. It’s something we’ve been calling “use case fever.” They’ve observed employees in many companies — almost in a frenzy — randomly creating voluminous amounts of use cases.
But when these same employees tried to implement them, they were unable to get the support needed because of inadequately defined benefits, the potential risks, and the lack of alignment with the company’s business imperatives. This randomness can deplete valuable resources, time, money and effort.
This doesn’t mean stop dreaming or being creative with use cases, but it does mean at some point you have to run these through an analysis on what the use case is intended to do, what value it generates, what risks it creates or avoids, and much more.
GenAI Pearls: Key Actions to Ensure Success
Many companies have reported in the past, and experience shows, that technology accounts for only about 30% of the success of digital transformation. The technology itself is essential but is not enough for transformation success. The other 70% is accounted for by business decisions and organizational, talent and change management actions.
What Digital Transformation Requires:
Technology is just part of the equation. Success comes down to a massive change management effort across many areas of the business, including talent.
With generative artificial intelligence, because of its creative nature in generating unique solutions, it will be even more important to focus on not only the technology, but the business imperatives, and the resulting organization, talent, and change issues to ensure the transformation is successful.
Following these five recommendations will help banks with focusing on a few key actions to increase their success rate with generative AI:
1. Set up a separate dedicated governance structure for generative AI.
A dedicated AI governance board, focused on GenAI initiatives, can expedite decision-making while addressing critical concerns such as algorithmic bias, risk, customer data and ethical implications. This board must work with existing banking governance structures, ensuring a harmonized, regulatory-compliant approach to GenAI implementation. One caution: Don’t make this a bureaucratic function that slows down the right solutions.
2. Run each of your potential use cases through an objective ‘use case’ scoring model.
Some use cases can be incredibly innovative and even potentially blockbuster ideas for a bank; but run it through the “wash” a few times and see if it holds up in terms of the value, risk tolerance, and business imperatives needed. Scoring models should focus on evaluating use cases for risk and compliance, revenue growth potential, operational efficiency and other important areas. This helps ensure your GenAI solutions will balance risk with reward.
3. Develop an actionable Generative AI NorthStar Strategy.
Clearly determine your business imperatives and guiding principles. This should include an understanding of the technology: what it can do, possibly do, and not do. Determine what generative AI will be and will not be at your company or in your function. Then develop actions to guide you to that NorthStar to move the process forward — together.
Want to go deep on AI best practices for banks?
Attend our AI Masterclass — Unlocking the Power of Artificial Intelligence in Banking — at The Financial Brand Forum 2024 on May 20-22 in Las Vegas. Led by Ron Shevlin, chief research officer at Cornerstone, this three-hour workshop will be jam-packed with lessons learned from industry leaders and real-world case studies.
For more information and to register, check out the Forum website.
4. Build prompt engineering skills in everyone who will build, leverage, or consume generative AI.
Prompt engineering as a job may or may not eventually go away, ironically because of AI, but the skills of prompt engineering will be an essential capability for anyone who builds these tools, leverages, or consumes them.
Prompt engineering skills help people think critically and creatively about problems, challenges, or opportunities in their work. Knowing how to ask the best questions, recognizing risk areas in generative AI, and understanding, at a visceral and intellectual level, what it can and cannot do builds talent that is more agile, thoughtful and risk-aware in executing work — all skills every bank needs more of.
5. Address the gaps between your current and needed organizational DNA.
Generative AI changes your organization’s DNA. This includes processes, workflow, governance, policies, structure, capabilities, talent, leadership and much more. Most believe the speed and breadth of change will be beyond anything we’ve ever experienced.
The DNA at most banks is not designed for what we are experiencing — and certainly not for what we are about to experience. It’s critical to determine proactively whether your organization’s DNA will enable or block generative AI. Blocking this technology from becoming part of your DNA means it also cannot become part of your organization’s value and success story.
- 3 Strategies for Enterprise AI Success That Are Tried and ‘Truist’
- How Leaders Can Ensure an Effective Change Management Effort
Implementing Generative AI at Your Bank
The promise of generative artificial intelligence in banking is undeniable, yet the stakes of haphazard implementation are high. It’s essential to avoid the pitfalls of random acts of digital through a well-thought-out, strategic approach. Start by establishing a dedicated GenAI governance board that collaboratively works within the existing governance structure, ensuring rapid but responsible decision-making.
“The promise of generative artificial intelligence in banking is undeniable, yet the stakes of haphazard implementation are high.”
Leverage an objective scoring model for potential use cases to align with your bank’s strategic imperatives. Develop a realistic GenAI NorthStar Strategy, incorporating clear business goals and explaining how this technology will contribute to enabling those goals. Prioritize the development of prompt engineering skills across teams companywide, preparing them to build, leverage and consume GenAI effectively.
Lastly, take steps to bridge gaps in your organization’s current and needed DNA, considering the profound changes GenAI brings to your workflows, governance and talent needs.
By taking these actions, banks can harness the transformative power of GenAI without falling into the trap of random, unstructured approaches to adoption. This is the time to plan how your bank will maximize the great potential that generative AI can bring.
About the author:
Garth Andrus is the president of Cognixia, an Ascendion company, coauthor of “The Technology Fallacy” from MIT Press, and a former board member at Deloitte Consulting.