4 Questions Every Bank CEO Should Ask Before Investing in Generative AI

Generative AI is taking the world by storm. Already, 45 percent of executives say they're ramping up their AI investments – just months after ChatGPT became a household name. 

For bank leaders racing to cash in on this technology, it's essential to understand its specific value, likely implementation timeline, and organization-wide impact. Here, I'll offer four questions to answer before investing in generative AI.

1. What specific value will generative AI bring to my bank?

Generative AI promises to deliver two primary sources of value: hyper-efficiency and hyper-personalization.

Let's look at an efficiency example first. Employees can use an AI-powered conversational chatbot to accelerate and automate traditionally manual tasks. Here's how it works:

1. A worker types in a query using natural language (e.g., "Give me a chart based on the following information.").

2. The AI, trained on relevant industry and organizational data, generates an output. Based on the worker's query, this output could be text, audio, or visual content.

3. The worker can add follow-up queries to refine the AI's output; the AI will remember all previous messages within a given conversation.

Generative AI technology can potentially create efficiencies in practically every role at a bank. For instance, consider these three use cases:

1. Analyst Copilot. Analysts can use generative AI to analyze financial data and write summaries of key findings or spoof fraudulent transactions to train fraud detection models.  

2. Marketing Assistant. Marketing professionals can ask an AI chatbot to segment customer data and suggest ways to target pain points for each group. They might also use generative AI to produce first drafts of marketing copy or mock-up visuals for a campaign.

3. Software Development. Developers can use generative AI to write or augment code, suggest improvements, produce documentation, and automate testing. This approach can fast-track things like application development and website updates.

Note that these examples relate to internal operations, but perhaps the most typical application: is customer-facing chatbots used for customer self-service. For example, customers can ask an AI-powered chatbot for help with their account login or surprise overdraft fees. Customers don't have to deal with long wait times; staff only take calls for the most complex issues.

Alongside these efficiency gains, generative AI will allow banks to hyper-personalize the customer experience.

For instance, a bank might A / B test specific card offers and product descriptions to learn what resonates with specific customer segments. But with generative AI, banks can generate the code to test near-infinite combinations and pinpoint the winners.

Other use cases include a personal AI banker for high-net-worth clients who proactively suggest financial products tailored to their needs. When it comes to generative AI and personalization, the possibilities are endless.

2. How long will it take to get a custom generative AI solution?

This answer will vary for every vendor, but as a rule of thumb: find a vendor that will deliver proofs of concept (POCs) in weeks, not months. And it's equally important to ensure each POC targets real problems at your bank. Otherwise, it's a wasted investment.

The right vendor will focus each POC for generative AI around a single problem (e.g., improving customer service response time and quality for a frequent, time-consuming issue). They'll run tests in a way that minimizes disruptions to the team. In a few weeks (no more than a month), they'll present results to help you decide if it makes sense to scale. If it doesn't, they'll develop another POC around the next small problem – and continue until you get the desired results.

This approach ensures you'll get the maximum ROI for your generative AI investment. And you won't have to wait years to see the impact.

3. How quickly should I scale AI?

In my experience, taking an incremental approach to scaling AI is best. The reason? Any generative AI application you embrace will impact every corner of your business. As you grow your applications, you'll need to measure how they affect…

Team configurations. As AI creates more efficiencies, you'll want to decide how much to staff teams. Maybe you will downsize your customer support team but use generative AI to boost capacity to handle customer issues. Or you keep the same number of software developers and use generative AI to turbocharge app development.

Leadership models. Team leaders and division heads will manage more AI workers but fewer total employees, affecting individual leadership placements and responsibilities throughout your organization.

Knowledge management. Although individual roles may shift, you'll need experienced employees to troubleshoot AI-generated outputs. For instance, if AI-generated code causes problems in your mobile banking app, you'll need a skilled software developer to fix the problem. And as you increase the amount of AI-generated outputs, you'll likely need more workers with AI skills.

Responsible use policies. There are serious concerns about generative AI's ability to handle sensitive data and generate accurate recommendations securely. Any successful org-wide rollout will depend on clear policies that place explicit boundaries around AI use. Such policies will also need to adapt as AI regulations take shape.

Understanding these risks is vital before going all-in on generative AI. I suggest using a team as a generative AI sandbox to start. Measure how it impacts their day-to-day work. Then, use these learnings to inform how you scale this technology throughout your organization. This way, you can maximize the enterprise-level efficiency gains from generative AI.

4. How can I secure buy-in for generative AI throughout my organization?

When talking to other leaders at your bank (e.g., the board of directors or the rest of the C-suite), I suggest focusing on the…

Proofs of concept (POC). Drawing on the POCs you've developed with your vendor, give concrete use cases for generative AI on, for example, the marketing or application development team.

Growth opportunities. Identify areas where generative AI can help you create a new revenue stream or fast-track your entrance into a new market.

But the biggest hurdle will be establishing trust in generative AI. Many workers are anxious about whether AI will replace them. And they may be wary of using a tool for their benefit today – only to be out of a job tomorrow.

In these conversations, make sure to lead with empathy. Highlight your team's indispensable value as knowledge workers: as AI powers more operations, they'll be critical to training models, overseeing execution, and ensuring quality outputs. Five or ten years down the line, I believe workers and AI will function as collaborators – not opponents – to help banks and their customers thrive.

Generative AI Isn't the Next Crypto

Generative AI has captured imaginations worldwide in its first year in the public domain. But so did the blockchain. And crypto. And the metaverse.

So what makes this technology different? Put simply, it's already changing the way people work. And it's happening in real-time.

For bank leaders, the question of when to invest is a matter of weeks and months rather than years. And with the right vendor, you can maximize your ROI to stay ahead of the competition.

---