Enterprise transformation is not simply about replacing legacy systems or adopting the latest technologies. It requires leaders to make deliberate decisions about what to retain, where to integrate modern capabilities, and when existing platforms have become barriers to growth.

In this PIECES interview, Subba Kethu, Head of Digital Technology at Kia North America, shares his approach to value-driven modernization, the foundations organizations need before scaling enterprise AI, and the leadership principles required to translate ambitious technology strategies into measurable business outcomes across regions and functions.

Q1. Many enterprises face pressure to modernize quickly while continuing to rely on stable legacy systems. What principles should leaders use when deciding what to retain, integrate, or replace?

I use a value-driven framework based on three questions: does this system create business value today, is it technically viable, and what's the risk-versus-reward of changing it? Modernization isn't a technology project — it's a business decision and treating it that way changes what you optimize for. The goal isn't to replace as much as possible; it's to maximize business value while reducing risk and increasing agility.
Here's the decision framework I use:

a. Retain what differentiates the business and still performs well.

If a system runs critical operations reliably, holds proprietary knowledge, meets performance and security requirements, and costs a reasonable amount to maintain — keep it. The test is simple: does this system create business value today without materially limiting our future? If yes, retain it.

b. Integrate when the core works but modern capabilities are needed.

Preserve stable transaction engines while exposing them through APIs, event-driven architecture, and cloud-based services. This lets you add AI, digital customer experiences, mobile, analytics, and automation without disrupting systems that already work.

c. Replace when the system becomes a strategic constraint.

Replace when a platform blocks growth, can't support new customer experiences, lacks required security, depends on disappearing skills, or simply costs more to maintain than modern alternatives. This is especially true when the business model has shifted or a SaaS platform offers an immediate strategic advantage — legacy CRM to Salesforce, legacy ERP to SAP S/4HANA, or custom-built applications to SaaS.

In large enterprises — including automotive organizations running ERP, dealer, manufacturing, and customer platforms — the winning strategy is rarely "replace everything." It's retaining what provides stability, integrating where innovation is needed, and replacing only where the legacy platform becomes a barrier to transformation.

2. As enterprise AI moves from isolated pilots into core business operations, what foundations must organizations establish across data, governance, cybersecurity, and accountability before scaling it?

The challenge is no longer proving that AI works. The challenge is building an organizational foundation that lets AI scale safely, consistently, and responsibly. AI readiness is primarily a governance and data challenge — not a technology challenge.

Data.

Before we consolidated 900+ data sets into a single enterprise data warehouse, teams across the business were often making decisions off different versions of the same numbers. That's the real cost of poor data — not an abstract risk, but decisions made on inconsistent ground. Data readiness and enterprise knowledge management must come before you scale AI usage, not after.

Governance.

We built governance at Kia to be the thing that lets us move faster with AI, not slower — covering privacy, accountability, transparency, security, lifecycle oversight, risk controls, and value measurement. When we deployed our RAG-based generative AI pipeline, having that structure already in place is what let us move from pilot to production without re-litigating security and access questions at every step.

Cybersecurity.

AI is a new attack surface, and it must be treated as one. Before scaling, organizations need data protection controls — encryption, access controls, data classification, usage policies — and AI-specific security controls: model security testing, prompt injection protection, agent monitoring, and third-party AI vendor assessments.

Accountability.

Someone must own the outcome, not just the output. That means clearly assigning responsibility for governance, incident response, lifecycle management, explainability, and oversight — before an incident force the question.

3. You have led technology initiatives and global teams across several countries. What have you learned about translating an ambitious technology strategy into measurable business outcomes across different regions and functions?

One of the biggest lessons I've learned is that a technology strategy only succeeds when it's translated into business language, business outcomes, and business ownership.

The first lesson: successful technology strategies start with business outcomes, not technology roadmaps. Whether it's cloud migration, ERP modernization, AI adoption, cybersecurity, or digital platforms, the first question should always be — what business problem are we solving, and how will we measure success?

The second: global alignment doesn't mean centralized control.

The most effective model is to govern strategically and execute locally. Different regions have different regulatory requirements, market conditions, and operational needs. A common governance framework, architecture standards, and cybersecurity controls create consistency, while regional teams keep enough flexibility to address local demands. That balance is what gives you scale without losing agility.

Third: large transformation programs succeed when they're broken into measurable increments. It reduces risk, demonstrates progress, builds stakeholder confidence, and lets you adjust priorities as conditions change.

The fourth: governance isn't a constraint on innovation — done well, it's what lets innovation scale.

As organizations increase investment in AI, data, and digital platforms, success depends on clear ownership, accountability, and measurable outcomes. The organizations that scale consistently are the ones that treat governance as a business enabler, not an administrative process.

And finally — transformation only becomes sustainable when people embrace it.

When business leaders, technology teams, and operational teams share ownership of outcomes, transformation moves beyond project implementation and becomes how the organization actually operates.

Ultimately, successful digital transformation isn't about deploying new technology.

It's about creating a clear line between strategy, governance, execution, and measurable business value. When those are aligned, technology becomes a driver of growth, resilience, and competitive advantage across every region and function.

Subba Kethu’s perspective makes one principle clear: successful digital transformation begins with business value, not technology deployment.

Modernization requires a balanced approach—preserving systems that continue to deliver value, integrating new capabilities where agility is needed, and replacing platforms only when they become strategic constraints. At the same time, AI can scale sustainably only when organizations establish reliable data, clear governance, strong cybersecurity controls, and accountable ownership.

Ultimately, technology becomes a meaningful driver of growth and resilience when strategy, governance, execution, and measurable outcomes are aligned—and when business and technology teams share responsibility for the results.

Subba Kethu, Head of Digital Technology, Kia North America

Reach the people shaping where capital goes next.