Expert Mode: From Mainframe to Modern CX: Unlocking Agility with AI-Augmented Modernization
This article was based on the interview with Unum CIO & CDO Shelia Anderson on AI-augmented enterprise modernization by Greg Kihlström, AI and MarTech keynote speaker for The Agile Brand with Greg Kihlström podcast. Listen to the original episode here:
For many enterprise marketing leaders, the words “mainframe modernization” trigger a familiar, low-grade sense of dread. It represents the monolithic, often inscrutable legacy systems that stand between your ambitious customer experience strategies and their real-world execution. This is the technical debt that slows down product launches, complicates personalization efforts, and makes true omnichannel orchestration feel more like a theoretical exercise than an achievable goal. You have the vision, the data, and the customer mandate, but progress is often bottlenecked by systems written in languages that predate the internet itself. It’s the classic story of “COBOL meets Cloud,” a challenge that can feel so immense it’s difficult to know where to even begin.
Yet, tackling this challenge is no longer optional; it’s the critical foundation for future growth and competitive differentiation. The real question isn’t if you should modernize, but how. How do you embark on a multi-year technological transformation while the business is, as they say, flying the plane? More importantly, how do you ensure these massive investments translate directly into tangible value for your customers and your business? We had the opportunity to explore this very topic at PegaWorld 2026 with Shelia Anderson, the EVP, Chief Information and Digital Officer at Unum, a 175-year-old market leader in the employee benefits space. Her pragmatic, business-first approach offers a compelling playbook for any leader staring down the barrel of their own modernization initiative, proving that even the most established enterprises can leapfrog into the future.
Start with Business Value, Not Technology
One of the most common pitfalls of any large-scale technology project is allowing it to become just that—a technology project. When initiatives are defined and measured by IT metrics alone, they can easily lose their connection to the strategic business outcomes they were meant to enable. For Anderson and her team at Unum, the starting point wasn’t a discussion about exiting the mainframe; it was about solving for the biggest business and customer challenges. This required a fundamental mindset shift across the organization, moving from a technology-first approach to one rooted in business-aligned value creation.
“It starts with our business, focusing on making sure that we’re solving for the biggest opportunity business challenges. And so, that… that means really looking at how we prioritize the work that we’re doing, saying no to the things that aren’t gonna yield the greatest value in our business, so that’s a bit of a new muscle for the company too.”
This is a powerful concept for marketing leaders. How often has a promising MarTech initiative been stymied because the underlying data or process is locked in a legacy system? By reframing the conversation around “value streams”—with clear business ownership from front to back—the technology investment becomes a direct enabler of a desired outcome. For Unum, this meant focusing on the claims process. For a marketing organization, it might mean focusing on the customer onboarding journey or the cross-sell/upsell process. When the business outcome (e.g., improved customer satisfaction, higher lifetime value) is the North Star, it becomes much easier to prioritize the necessary technological work and, crucially, to say “no” to projects that don’t directly support it. This alignment ensures that IT isn’t just modernizing for modernization’s sake; they’re building the specific capabilities marketing and product teams need to win.
Demystifying “AI-Ready”
“AI-ready” has quickly become one of the most overused and abstract phrases in the corporate lexicon. It can feel daunting, suggesting a state of data purity and architectural perfection that few enterprises possess. Anderson, however, offers a refreshingly practical and bifurcated approach to making an organization truly AI-ready. It’s not a single, monolithic goal but a dual-track effort that addresses different needs and skill sets across the company. This approach makes the concept tangible and actionable, moving it from a buzzword to an operational reality.
“One is what I call everyday AI for the broad employee population… we’re actually encouraging citizen development throughout the enterprise… On the other side of that is what I like to call more of the customer embedded AI. Much of that does require more of a deep engineering skillset, and for example, inside of engineering… we actually established a goal, so that 100% of my engineering staff would be fully trained in leveraging AI in their daily work.”
This two-pronged strategy is directly applicable to any marketing department. “Everyday AI” is about empowerment and productivity. It’s about giving your content creators, campaign managers, and analysts tools like Microsoft Copilot or other generative AI assistants to streamline their daily work, generate first drafts, and synthesize data more quickly. This track focuses on broad education, adoption, and taking the fear out of the technology.
“Customer-embedded AI,” on the other hand, is the heavy-lifting, strategic work that transforms the customer experience. This is the realm of personalization engines, next-best-action models, and predictive analytics that require deep engineering and data science expertise. By separating these two tracks, you can make immediate progress on employee productivity while simultaneously tackling the more complex, foundational work required for sophisticated, AI-driven marketing. One track provides immediate, tangible value, while the other builds the long-term competitive advantage.
Measure Business Outcomes, Not Just IT Activity
The final, and perhaps most critical, piece of the puzzle is measurement. A successful modernization project cannot be judged by traditional IT metrics like system uptime or lines of code rewritten. While those are important for the team doing the work, they mean very little to the C-suite or the marketing leader trying to improve customer retention. True success must be measured in the language of the business. Anderson emphasizes a tight coupling between technology investments and the operational outcomes they influence, ensuring that value is not just promised, but proven.
“Gone are the days of simply reporting the internal IT metrics. We still do those so that we can see improved productivity on an agile squad, for example, but the ultimate outcome is the outcome that you’re driving for the business… in the claims journey for us, it’s customer satisfaction, average handle time, for example.”
This is the alignment that marketing leaders dream of. When your CIO is reporting on Customer Satisfaction scores as a key metric for a platform modernization project, you know you have a true strategic partner. This shift requires building cross-functional teams and shared KPIs from the outset. Before a single server is provisioned, the business, operations, and technology teams must agree on what success looks like. Will it be a reduction in call handle time? An increase in NPS? A faster speed-to-market for new product variations? By defining and tracking these business-level outcomes, the technology investment is held accountable to delivering real, measurable value, creating a virtuous cycle of trust and further investment.
The Courage to Begin
The sheer scale of enterprise modernization can induce a state of analysis paralysis. Faced with decades of accumulated complexity, the temptation is to wait for the perfect plan, the perfect budget, and the perfect moment. Anderson’s most resonant piece of advice cuts directly through this inertia: just get moving. Progress over perfection is the mantra for the modern enterprise, and it’s a principle that agile marketers know well.
“My biggest piece of advice is just get moving. So put your toe in the water before you know that you can swim, for example. This is one of those things that you’re going to have to take a test and learn approach to… It doesn’t have to be perfect.”
This is not a call for recklessness, but for focused, iterative progress. Unum didn’t try to boil the ocean; they started with a specific, high-value part of the business—the claims process—to prove out their methods. For marketing, this could mean modernizing the systems that support a single customer journey or a specific line of business. By delivering value in focused increments, you build confidence, secure buy-in, and generate the momentum needed to tackle the next challenge. This “peel the onion” approach allows you to learn and adapt, de-risking the overall transformation while continuously delivering value to the business.
The journey from a COBOL-based mainframe to a modern, AI-augmented, cloud-native architecture is undeniably complex. However, as Shelia Anderson’s experience at Unum demonstrates, it is not an insurmountable IT problem. It is a business transformation, one that must be led with a clear focus on customer and business value from day one. By prioritizing relentlessly, making AI a practical tool for everyone, measuring what truly matters, and having the courage to simply begin, leaders can dismantle the legacy barriers that have held them back.
For marketing leaders, this isn’t just about cheering from the sidelines. It’s about being an active partner in the process, helping to define the value streams, co-owning the business outcomes, and championing the iterative approach. The end goal isn’t just a new tech stack; it’s organizational agility. It’s the ability to launch, test, and learn at the speed of the market. The ultimate prize of modernization is not the absence of legacy technology, but the presence of an enterprise-wide capability for continuous, customer-centric innovation. And that is a future worth investing in.
