Technology Transformation & The Strategic C-Suite
Roundtable
November 2, 2023 | New York
Companies across industries are investing in integrations between finance and non-finance data to meet the demands of an evolving technology landscape. Executives from finance, technology and operations gathered for this roundtable dinner to share their experience and ideas about how their firms pair a transformational mindset with the tools needed to create robust, efficient and effective outcomes throughout their organizations.
Participants:
-
- Bjorn Austraat, Senior Vice President, Head of AI Acceleration, Truist
- Oded Berson, Senior Vice President, Data & Analytics Enablement, Chubb
- Mathias Carlbaum, President & CEO, Navistar
- Nadia Carta, Head of Data Transformation, Google
- Lauren Davis, Associate Partner, Global Finance Transformation, IBM Consulting
- Jay DePaul, Chief Cybersecurity & Technology Risk Officer, Dun & Bradstreet
- Rafi Ezry, Managing Partner, US Industrial Market, IBM Consulting
- Iwao Fusillo, Global Head of Data & Analytics, eCommerce, PepsiCo
- Karen Hanlon, Executive Vice President & Chief Operating Officer, Highmark Health
- Yoana Land, CFO Transformation, L’Oreal North America
- Dr. Arezu Moghadam, Managing Director, Global Head of Data Science, JP Morgan Chase & Co.
- Jonathan Nash, Managing Director, Head of Investment Risk, Voya Investment Management
- Scott Roe, Chief Financial Officer & Chief Operating Officer, Tapestry
Bloomberg Participants:
- Lisa Mateo, Business Correspondent, Bloomberg
- Lynn Doan, Managing Editor, Europe & US East Technology & Global Cybersecurity, Bloomberg
Roundtable Highlights
After introductions, Lisa Mateo, Business Correspondent, Bloomberg asked participants to chat about using AI to manage large amounts of data and automate decision-making.
Lauren Davis, Associate Partner, Global Finance Transformation, IBM Consulting advocated for establishing a governing board “that not only certifies what you’re going to govern and what you’re going to put in your master data management pool or leave unstructured, but also the kind of data feeding through your AI and large language products.” There is a lot of focus there, because it’s foundational. She urged everyone to start committees and have these conversations.
Responding to that, Scott Roe, Chief Financial Officer & Chief Operating Officer, Tapestry said “the lawyer side” of it and compliance are the hardest part in his organization. “The world is moving faster than the law, and we’re really struggling sometimes to figure out the balance between moving fast and empowering people, and how you use the data and push it down, versus this need to control it.”
Mathias Carlbaum, President & CEO, Navistar jumped into a conversation about job skills. In the public schools his children attend in the U.S., interest in moving toward this kind of technology is not promoted. “ChatGPT came out, the first thing was, let’s check over the shoulder of the kids. That was the reaction; they’re cheating. It will take many years until the system learns how to use these tools to form education to create value from their reality.” He wondered if anyone knows of private-public initiatives to target younger students to create the big shift needed in education.
Roe responded that he was on a community college board that did a lot of partnerships like aircraft mechanics and high-end automotive; highly specialized skill sets. “They were sponsored by the companies, so you come in and they make you an aircraft mechanic, and you can make $80,000 a year with no college education.” He believes that will become a more widespread approach. At Tapestry, they did this in cybersecurity. “We couldn’t find people, and the ones we had were so expensive, and they kept getting poached. So, we grew our own.”
Thinking in a broader scope, Carlbaum mentioned studies that show the “cloud-born” generation, while immersed in technology like social media, are not exposed to it in terms of data and ways to create value from that experience.
When it comes to data, not everybody knows that large language models are trained on massive amounts of data and many are closed systems, said Rafi Ezry, Managing Partner, US Industrial Market, IBM Consulting. It results in lost data and unknown sources, copyrights and other elements that are going into models. “Data lineage is something that companies need to look at, otherwise they add to their liability profile by using those models.”
During another conversation, Ezry said, “I think the challenge now is to deliver scale on custom software development. Every company around the table will eventually have a significant custom software stack that they own and that’s part of their digital product portfolio and intellectual property.” As that scale does not yet exist, some companies are building their own. Others need an affordable solution. “There has to be deliberate thought on how to kind of bridge that gap.”
Nadia Carta, Head of Data Transformation, Google added to those thoughts, reiterating Ezry’s point during the earlier panel discussion, ‘Once the data is in, it’s in.’ “I think too many people right now think it’s almost like a game, thinking, I’m going to put inside ChatGPT or Bard content like my flow chart, and then they don’t think that once you put it there it’s actually accessible to everyone.” On copywriting, “I don’t even know how we’re going to solve it because it’s actually very difficult to validate, unless technologies find a way, like watermarks.”
Intrigued by Karen Hanlon’s earlier description of her company’s advisory board, and taking a similar approach to addressing new technology, Carta asked where they find qualified people. Hanlon, Executive Vice President & Chief Operating Officer, Highmark Health, explained that it started with the realization that governing board members typically have long tenures and stronger historical expertise, prompting the need for forward-looking perspectives on emerging technologies. They looked for recommendations from company and university leaders, Google Cloud, vendors, and those who are publishing. “It’s still very much in its infancy, but it has promoted good dialogue. In fact, we had an entire discussion with them last week on the topic of engagement and what tools are they seeing out there.”
When companies turn to AI to enhance their campaigns, it presents an ethical scenario. Yoana Land, CFO Transformation, L’Oreal North America used the example of an outsourced lipstick campaign, where they were initially impressed with the image results; better than what L’Oreal’s in-house team could do. Then they realized the eyes and mouth were taken from three different women. “Who has the rights over that? There’s multiple data sources.” The company opted not to use it, viewing it as deceptive to the consumer. But the technology is still there for anyone to use, she said, making it a question of data privacy.
Jay DePaul, Chief Cybersecurity & Technology Risk Officer, Dun & Bradstreet brought up an aspect that didn’t come up during afternoon sessions, noting that it is important to keep in mind licensing and third parties that are creating content with Generative AI. “Microsoft Office and others that are using more content in their products, they’re all turning on CoPilot or some other capability. Everybody’s turning on some form of AI, and we now have the capability to click through, and now they’ve got our licensing. Be super careful about that.”
Daring to be geeky, Iwao Fusillo, Global Head of Data & Analytics, eCommerce, PepsiCo described how Pepsico is using Generative AI to vectorize and categorize images to gain insight into what will resonate best with consumers. “We can very quickly understand, when we’re selling muscle milk, is it better to show the product or is it better to show a pasture, or is it better to show people smiling. We weren’t able to do that before AI because humans can’t scale.” He noted they use a very straightforward, free computer vision tool from Amazon.
Plaguing the consumer packaged goods industry, Fusillo said, is “ridiculously” unstructured data and a product catalog full of codes that are like hieroglyphics. “Even before the days of Gen AI, we built AI ML engines to structure it.” He predicted “very interesting” monetization and cost-defraying opportunities are ahead.
On that topic, Dr. Arezu Moghadam, Managing Director, Global Head of Data Science, JP Morgan Chase & Co. described how they are optimizing investing. Agents who previously had to stay on top of all market trends now get an assist from gen AI as it listens in on conversations with clients. “It picks up on keywords, product mentions and tickers. It pops up information for them quickly so that conversation goes much smoother and they sound much smarter.” Results also show an increase in the number of tickets they are closing. AI is also used to fine-tune stock selections to match client “themes” by processing all the news and other information about specific companies. How does that impact jobs? “I believe that we are very, very far away from putting all of this onto pilots. Accountability for us is still with the human. I don’t think AI would replace a portfolio manager, but a portfolio manager who knows how to use AI is going to replace the portfolio manager who doesn’t know how to use AI.”
In the insurance industry, large, unstructured data, in click nodes or underwriting guidelines, make people more efficient, Oded Berson, Senior Vice President, Data & Analytics Enablement, Chubb said. “Gradually, it’s probably going to move towards taking the simpler types of policies, or maybe you’re not going to do it for a new policy but for a renewal, where you know more about the client. If gen AI signs off on it, it could be a big efficiency. In insurance, and in manufacturing and other industries, too, there is big, big pressure on expense ratios.”
This Bloomberg briefing was Proudly Sponsored By
——————————
Join the Conversation: #TechTransformation
Instagram: @BloombergLive
LinkedIn: Bloomberg Live
Twitter: @BloombergLive
Interested in more Bloomberg Live virtual events? Sign up here to get alerts.
——————————