How 4 CEOs are approaching generative AI use cases in their companies

By Larry Dignan

Generative AI is the question of the day on many earnings conference calls, but the approach to leveraging the technology depends on the industry. Here's a look at how Bank of America, Airbnb, Lemonade and Deutsche Telekom are thinking about generative AI.

Bank of America CEO Brian Moynihan: Generative AI needs more transparency

With existing investments in its ERICA natural language AI, CEO Brian Moynihan needs more transparency before fully leveraging ChatGPT and other generative AI technologies. That approach shouldn't be too surprising given banks are highly regulated.

Speaking on Bank of America's first quarter earnings conference call in April, Moynihan said the bank started working on ERICA, a predictive Q&A bot, years ago. The bank created its own language for ERICA and deployed it. The catch is that ERICA was only looking at Bank of America systems without third party data.

"Because ERICA is captive to our data, where it's just looking at our systems, finding information and giving to clients it is really a service capability," explained Moynihan. "We've also taken ERICA internally and applied it to help us do work and we've seen it have those benefits."

He added generative AI can improve computer coding productivity and content creation as well as analytics. However, Moynihan said transparency is going to be an issue if you're integrating your corporate data with generative AI.

Research: Microsoft Shifts Coding From Typing to Language With ChatGPT

"The reason why a lot of it stopped in our industry and other industries was, it wasn't clear how it worked. It was your data and the outside world's data and how it would interact and pull stuff out and we have to be careful with that," he said. "We have to understand how the decisions are made, and, frankly, follow the laws and rules and regulations on lending."

"We understand the value of it, but we will carefully apply it and we see a great value. I don't think it's a great value in the next month, but in the overall sense, it will help us continue to manage the headcount down which we've been doing this quarter. This management team started with this company in 2010 with 285,000 and 300,000 people working here and we're running the same size company with 216,000 people or bigger company doing more stuff and that's been aided by digitization."

Airbnb, CEO Brian Chesky: All about the tuning

Chesky said generative AI may be as big of a platform shift as the Internet was. The competitive advantage will be for the companies that take a base model and uniquely tune it.

Chesky said on Airbnb's first quarter earnings call:

"All of this is going to be built on the base model. The base models, the large language models, are like major infrastructure investments. Some of these models might cost tens of billions of dollars towards the compute power. And so, think of that as essentially like building a highway. It's a major infrastructure project. And we're not going to do that. We're not an infrastructure company. But we're going to build the cars on the highway. On top of the base model is the tuning of the model."

He said the tuning of the model will be based on the customer data you have. When the base model is combined with customer data then there's a lot of innovation.

"I think that going forward, Airbnb is going to be pretty different. Instead of asking you questions like where are you going and when are you going, I want us to build a robust profile about you, learn more about you and ask you 2 bigger and more fundamental questions: who are you? And what do you want? Think of us with AI as building the ultimate AI concierge that could understand you."

MARKET OVERVIEW: Analytics and Business Intelligence Evolve for Cloud, Embedding, and Generative AI

Lemonade co-CEO Shai Wininger: Integration is easy if you have existing AI stack

Lemonade is a digital insurance company that has been using machine learning and AI for years. It uses models to predict losses and risk down the zip code as well as each discrete home it insures. The goal for the company is to have true precision in pricing and underwriting.

"Lemonade was built as a tech-powered insurance company. We were the first to provide customers with human-like chat-based experience that works 24/7 and handles all of our direct sales," said Wininger, speaking on the company's first quarter earnings call. "On the back-end, we built a first of its kind insurance operating system that lets our team service our customers efficiently and delightfully. This year, ChatGPT 4 and other large language models made a huge leap forward. The potential impact this technology brings to businesses like ours is substantial."

Since generative AI can reason and learn it has the potential to improve efficiency and customer service as well as risk assessment and underwriting decisions. "This technology will help us better anticipate customer needs, respond to more claims instantly and ultimately provide better coverage at lower costs," he said. "For our competitors, though, adapting to this change will not be easy. A traditional insurance company depends on hundreds of disparate software tools to run its business. many of which are outdated legacy systems built decades ago by third-party vendors."

Wininger's argument is that companies without an AI-friendly stack will be buried by technical debt. Lemonade's bet eight years ago revolved around conversational UI and chatbots and it has updated models as they have evolved. He said:

"Today, we use dozens of AI models to do pricing, underwriting, customer service, payments and many other internal operations. We even built our own internal AI framework to help us manage and deploy models seamlessly and quickly across the organization. In just a matter of days, our team was able to add ChatGPT and other generative AI tools to our platform. We now have more than hundred different initiatives, which we believe can have a meaningful impact on our business. As a result, we expect to see more savings in the next 18 months and anticipate continued improvements in both our expense and loss ratios."

Deutsche Telekom AG CEO Tim Hottges: Eying customer service benefits

Speaking on DT's earnings conference call, Hottges was inevitably asked about generative AI. He said:

"There was never in the last 20 years that I recall a technology that had so much impact on thinking in management but as well in the organization as the announcement of ChatGPT and OpenAI. This is really a game changer in the way we look at it."

DT launched a program that features a centralized team that works with OpenAI's ChatGPT as well as Large Language Models including fine tuning. "We are fine-tuning all the elements of it with our foundation in the customer service arena," he said. "We used it already in the past, but we will now trying to exponentially develop the area of chatbots, call center support, network optimization and other areas where big data is affected."

Generative AI guide: ChatGPT: Hype or the Future of Customer Experience