3 Questions to Ask About Your Corporate Strategy in the Age of AI

Today’s digital giants are changing the definition of a diversified portfolio. As industrial giants like GE and Honeywell—once models of diversified portfolios—split up, consider how companies like Amazon, Apple, and Alphabet operate across many unrelated sectors, like e-commerce, cloud computing, entertainment, financial services, smart homes, logistics, and healthcare. As traditional diversification theories based on economies of scale and scope become less relevant in the digital age, the growth of AI is also ushering in new opportunities to create value with shared data, expertise, and network effects, and a new imperative for business leaders to reshape their strategies.

The example of Intercorp Perú Ltd., a company we advised in Latin America, is illustrative. Carlos Rodriguez-Pastor, chair of Intercorp, was intrigued by AI’s transformative potential. His conglomerate has built brands across various sectors like financial services, retail, health, and education. The company started in Peru and was expanding into other Latin American markets. However, leaders faced a challenge common to many multi-business companies: While their decentralized structure was effective in supporting independent operations, it also limited the company’s ability to systematically use data and AI for cross-selling, upselling, and personalized offerings to its 17 million customers.

Intercorp’s story shows an evolving narrative that should inspire others who are rethinking their corporate strategy. In the chair’s conversations with us, he asked, “What’s the future of Intercorp in the age of AI?” We distilled that core question into three parts to help Intercorp develop its roadmap.

Transitioning from an industrial-era conglomerate to an AI-driven corporate model requires answering the three questions in the correct order. Like constructing a house, you must first secure the foundation (portfolio logic), then design the structure (operating model), and only then install the advanced systems (algorithms). As a corporate strategist, your challenge is to generate equally compelling answers to these three questions:

1. What is our portfolio logic?

Many conglomerates are poorly configured for an AI-driven world. The question isn’t how diversified you are, but whether your businesses would benefit from shared data, expertise, and network effects. Clarify how data flows across businesses to uncover insights that no single unit could generate on its own and how these flows reinforce both corporate learning and competitive distinctiveness.

We first asked Intercorp leaders to define the logic that would connect their divisions into a unified portfolio. We advised that the logic shouldn’t focus solely on operational efficiency, market dominance, or managing business cycles, but rather on how their different units contribute to and benefit from shared learning and capability building, supported by data on their operations and customer interactions.

The leadership team grappled with a fundamental question: Why should its diverse set of businesses remain under a single corporate umbrella? Instead of settling on sharing physical assets or financial resources as reason enough, we helped them uncover a deeper insight: the potential to leverage AI and data to boost financial inclusion and middle-class services for Peru’s citizens. In our workshops with Intercorp leaders, we drew on ideas from our 2022 HBR article on the power of datagraphs and invited participants to envision how Intercorp could deliver truly personalized recommendations by connecting the full breadth of customer interactions across its businesses.

It became clear almost immediately that each unit was operating with a narrow, fragmented view of the customer. By linking data across their financial services, retail, health, and education businesses, they saw how much more effectively they could anticipate needs, coordinate offerings, and unlock value for their 17 million customers.

This stands in sharp contrast to the traditional reasons companies have assembled loosely linked portfolios, like shared physical assets, synchronized industry lifecycles, or the ability to move financial resources across units. That logic made sense in an industrial era grounded in tangible assets and predictable investment cycles, but they offer little advantage in an intelligence-driven economy. Shared plants or procurement budgets don’t improve decision quality, lifecycle alignment doesn’t help firms learn faster, and financial resource-sharing is easy to replicate when capital is abundant. What creates real, defensible advantage today is the ability to link data across businesses to generate richer customer understanding, sharper predictions, and more personalized engagement—an engine of value that traditional portfolios simply cannot match.

2. What is our corporate operating model?

Traditional divisional or matrix structures, built for control, are giving way to more fluid, cross-functional forms designed for speed and learning. Show how you’ve built systems that enable peer-to-peer collaboration, rapid experimentation, and fast diffusion of knowledge. The aim is to scale AI capabilities across the portfolio without stifling entrepreneurial energy.

We asked Intercorp’s leaders to identify ways to promote rapid experimentation and learning across divisions by leveraging data and AI. They recognized that their new operating model must strike a balance between divisional autonomy and shared knowledge and avoid the risks of full decentralization, which hampers knowledge sharing, and the opposite extreme of rigid centralization, which stifles innovation. We emphasized the importance of rethinking industrial-era hierarchical models that focused on control rather than collaboration. Guided by us, the team adopted a core principle: data and algorithms are company-wide assets that all business units can access and share.

Companies often default to a centralized digital unit that delivers infrastructure as a shared support function. But we encouraged Intercorp’s leaders to move beyond this mindset and see digital not as a service center, but as a peer-to-peer mechanism that links data across units—enabling collaboration, insight-sharing across experiments, and coordinated value creation that no centralized function alone can deliver. For example, we encouraged them to consider how their financial services unit, Interbank, could become a center of excellence for analytical models. When the bank develops new AI models for credit risk assessment, these capabilities shouldn’t stay confined. Instead, the core insights could be adapted to improve student performance prediction in Intercorp’s education division and to guide customer behavior analysis in their retail operations. This provided a tangible proof point of how cross-business learning would be institutionalized in practice.

This choice to invest in peer-to-peer collaboration rather than a centralized model directly stems from Intercorp’s overarching goal to serve the growing middle class across multiple touchpoints. Intercorp recognized that it could offer customers highly personalized options at scale and quickly—in ways that its businesses working independently couldn’t achieve and that their competitors couldn’t replicate. Payment data from Interbank helps improve their retail units, while insights into customer behavior from retail guide the development of educational services. These connections create a positive feedback loop of learning that would be very difficult to duplicate through market transactions. We also emphasized that Intercorp shouldn’t develop separate digital units that might not integrate well with the independent units within the company. Instead, they’ve seamlessly integrated data and analytics into the company’s operating models.

3. What are our unique algorithms?

Finally, demonstrate how proprietary data and domain expertise combine through algorithms to create enduring value. The goal isn’t just technical superiority, but cumulative learning—drawing from unique data sources, enriching them through partner integration, and leveraging these insights to create economies of expertise that grow over time.

We guided Intercorp leaders to focus on leveraging their proprietary data models and algorithms to secure a competitive advantage across their portfolio. Working with us, they understood that it’s time to move beyond traditional industrial-age skills, such as engine technology (e.g., Honda), miniaturization (Sony), storytelling (Disney), or brand-building (Virgin), to develop a distinctive algorithmic advantage by leveraging its rich proprietary data.

Unlike companies that rush to adopt AI solutions without organizational frameworks for coordination, Intercorp committed itself to developing its algorithmic capabilities collaboratively. We explained how they could create AI models that learn from their millions of interactions across 17 million customers to harness data network effects. They recognized that this approach could help them build advanced predictive abilities, leading to prescriptive recommendations across the financial services, retail, and education sectors. For example, Intercorp’s Health Platform is in a rapid growth phase, making bold, irreversible decisions about where to open new hospitals and which specialties to offer. By using algorithms to analyze extensive data from Intercorp’s pharmacies and credit card transactions, Health Platform can turn insights into smarter decisions for expansion.

The Sequence Matters

Intercorp’s experience shows that digital transformation doesn’t have to diminish the entrepreneurial spirit that makes conglomerates resilient. By following a disciplined sequence—from purpose to structure to technical capability—companies can create new forms of corporate advantage while maintaining agility. Carlos Rodriguez-Pastor acknowledges that the shift from an industrial-age to an intelligence-age conglomerate is iterative: Each stage reinforces the next.

. . .

There’s concern that we might be in an AI bubble, with unprecedented investments in AI infrastructure. These investments will only yield returns when AI capabilities are effectively integrated into the strategy and structure of corporations. Corporate strategy must shift from the industrial logic of scale and scope, as described by Alfred Chandler, to a modern intelligence logic centered on knowledge and expertise. The Intercorp case illustrates how a company can lay the groundwork for success in the era of AI.