Highlights from our recent Metis Strategy Summit are below. Check out our Youtube channel and Technovation podcast in the coming weeks for recordings of the conversations.
Organizations across industries are moving beyond initial AI experimentation, focusing on driving implementation, proving and measuring ROI, and developing the next generation of talent as they apply AI to a broader number of business challenges.
As multiple executives emphasized, strong data foundations are essential to any successful AI implementation. Marina Bellini, President of Global Business Services at Mars, noted that the hype around AI has led to more focus on ensuring those foundations are in place. “This is the dream of the CIO: that people will actually start working on data quality.”
This year has also seen increased focus on AI’s ability to deliver value. Augment CEO Scott Dietzen said 2024 “is the year where tech teams are looking for proof and return on investment,” something not always clear or easy to measure for software such as Copilot productivity tools.
Organizations are finding new and innovative ways to apply data and AI to business challenges. Royal Caribbean Group CIO Martha Poulter described how the company transformed traditional food service operations into data-driven processes. Initially, “you would order what you thought, cook what you thought, and serve what you thought. It was gut based,” she explained. By measuring proteins before and after cooking and analyzing everything from ordering to de-thawing to waste, Royal Caribbean was able to generate tens of millions in savings while improving sustainability. “You’d never think food can be an AI problem, but it is,” Poulter said.
Similarly, Avis Budget Group is using an AI-based modeling and prediction system to address asset utilization challenges and ensure cars are on the road for the greatest amount of time. Chief Digital & Innovation Officer Ravi Simhambhatla explained how the company is aiming to break through the 70% utilization ceiling for its vehicle fleet. “If you have physical assets that aren’t being utilized, it’s costing the company money,” he said. “We hit this glass ceiling and asked ourselves why can’t we go to 80% or 90%? It turns out it’s data.”
Technology leaders discussed various approaches to managing and organizing AI initiatives across their organizations. A common thread across nearly all of them was the importance of bringing together technology and business leaders to identify valuable use cases and deliver on them faster. NRG’s Chief Data and Technology Officer, Dak Liyanearachchi, talked about establishing a transformation office that bridges data, business, and technology teams. At Berkadia, an AI Council that includes both business partners and technology leaders drives deeper engagement and keeps discussions focused on value, Chief Information and Innovation Officer Damu Bashyam said.
As mentioned throughout the event, these new organizational structures place particular emphasis on modern technology stacks and data practices. Nicholas Parrotta, Chief Digital and Information Officer at HARMAN International, outlined the company’s evolution from infrastructure-as-a-service to data-as-a-service, and using that data to create more personalized experiences on wheels as the world moves toward autonomous vehicles. “We start with how we do the big stuff with architecture, then product, and now data and being able to drive those as revenue and capabilities,” he said.
Capital One CIO Rob Alexander detailed the company’s platform strategy, explaining how the organization built dedicated infrastructure for machine learning, feature engineering, and now generative AI applications. When it comes to AI, he noted that while it’s “easy to get 70% accuracy out of the box, all the work is getting from 70-75% accuracy, which involves training and fine tuning.” Being in a position to leverage AI today has been a 12-year journey for Capital One, Alexander said, one that has included transforming “everything about who we are” to become a successful technology company and a winner in the banking industry.
Leaders emphasized the need for pragmatic approaches to AI implementation. Mastercard CTO Ed McLaughlin noted three questions a review panel considers when evaluating the feasibility of a new AI initiative: “Does it work, is it worth doing, and does it align to our ethics?” If ChatGPT-style search responses are 10 times more expensive than traditional methods, for example, the question then is whether they can deliver 12 times more value or be that much more useful. McLaughlin underscored the need to assess both the right way to solve a particular problem and whether there are returns on the work being done.
Dietzen added that NPS and engineer satisfaction can also be indicative of value. “If you make engineers delighted, you’ll tend to do well in your organization,” he said.
Chris Davis, Partner and West Coast Office Lead at Metis Strategy, advises technology leaders to ensure that there is product management in every layer of the AI stack, including the application of AI to business processes, the marketplace of horizontal and reusable capabilities across use cases, and underlying foundational models and model development. Business value should be measured relative to components of the stack, especially with generative AI, Davis said.
Effective product management requires teams across the organization to sharpen their product mindset. Cigna’s Chief Digital & Analytics Officer Katya Andresen outlined three elements of that product mindset: identifying real problems for real users, validating through testing and learning, and unlocking value. She cautioned against common pitfalls like “death by a thousand pilots,” in which proofs of concept pile up and eventually become unmanageable. Organizational silos can present a challenge. “We find a lot of opportunities to streamline operations, but there has to be a really deep partnership across tech and ops,” she said. Otherwise, “tech gets upset that ops don’t use their products and ops says well what you gave us didn’t solve our problem.”
Organizations are rethinking their talent development strategies as the landscape evolves. That involves both upskilling internal talent and expanding talent pools across geographies. Land O’Lakes CTO Teddy Bekele described moving from a roughly 50-50 mix of in-house and external talent to a model in which contractors and third parties make up a more significant portion of the talent pool, taking on much of the development work while in-house employees lead the teams. The approach allows for increased flexibility in team sizes depending on shifting enterprise needs. The change was driven by three key factors: accessing expertise, maintaining flexibility to scale teams up or down, and increasing nimbleness.
Upskilling also remains a key focus. At FINRA, Chief Technology Engineering Officer Tigran Khrimian’s team is teaching developers generative AI skills and has seen demonstrable success with using natural language prompting to create “good code” for the company. “Developers with code assistant tools will replace developers who don’t use them,” he said.
Corning’s Chief Digital and Information Officer Soumya Seetharam detailed the company’s three-pronged approach to talent development: creating strategic digital and IT hubs around the world to ensure global talent access; launching a digital literacy program with dedicated “revitalization days” for learning rather than meetings; and expanding the talent pipeline through technology internship and rotational development programs globally. “In the future every person for every function will have some technology in their background,” she predicted.
Technology leadership roles are undergoing significant transformation, reflecting the strategic importance of technology in business operations. According to Katie Graham Shannon, global head of the Digital and Technology Officers Practice at Heidrick & Struggles, the traditional CIO title is becoming less common. Of 23 recent technology leader placements at Fortune 200 companies,18 did not have the CIO title, and 52% were newly created positions with expanded roles. She noted that there is also a shift in reporting structures, with more CIOs reporting to the CEO, and a greater focus on technology leaders’ ability to create and protect value and attract talent, among other responsibilities.
“If we could use the title ‘orchestrator’ it would make more sense,” Shannon added, explaining that today’s technology leaders create value and orchestrate initiatives across the entire C-suite. This expanded scope includes both customer-facing initiatives and internal operational efficiencies with “equal pressure and emphasis” in both areas.
The role is also becoming more business-oriented, particularly in relation to managing technologies like AI. “A properly formatted conversation about AI is not a tech conversation, it’s a business conversation,” observed Henry Man, Co-Founder and Managing Partner at Candela Search. This presents an opportunity for technology leaders to “have a seat at the table” when business colleagues might lean out of technical discussions.
That expanded purview extends to technology leaders on boards or seeking director positions. “There’s no market for a one-issue board member,” said Art Hopkins, who leads the Technology Officers Practice at Russell Reynolds. “You need to show business acumen and a P&L. Go to the CEO and say I’d like to be the executive sponsor of this new incubator. This is a solid step in that direction.”
