Ways CIOs can synergize modernization and legacy infrastructure

The old approach to dealing with legacy systems was simple: rip them out and start over. These days, though, most CIOs will tell you a different story that sometimes the best path forward isn’t about replacing everything but making what you have work better.

With emerging tech multiplying at such a furious pace, some companies are getting creative about how to retrofit old technology. Sometimes the choice to keep legacy systems in place comes down to if it’s still working, why change it? But the reality is rarely that clear cut.

Take Yahoo, for instance. The company’s physical storage infrastructure has been humming along for years, handling massive volumes of email data with rock-solid reliability. “We’ve kept it because it’s delivering the reliability and compliance we need,” says CTO Lee-Ming Zen. “We want to make sure we do the highest quality job, and having that on-premise physical storage lets us do that in a very safe, controlled way.”

Liberty Mutual’s mainframe processes billions of critical transactions every year, including policy management, claims processing, and other essential functions. Andrew Palmer, the company’s EVP and CIO, says it remains in place because of the resiliency it provides for essential insurance functions. “The decision isn’t about clinging to old technology,” he says. “It’s about maintaining stability for mission-critical operations.”

Industry specifics can play a huge role, too. Brown-Forman has been using the same product lifecycle management tool for over 15 years to track every recipe variation for its brands, including Jack Daniel’s and Woodford Reserve. Mike Homer, who heads up the firm’s master data management unit, points out that the organization’s entire liquid history lives in that system. “The tool also happens to be a common one used in the spirits industry, so we’re not the only ones using it,” he says.

The challenges can be even more pressing for smaller companies. At network management company Auvik, COO John Astorino faces the reality of maintaining a legacy system built in Scala. While Scala has proven reliable over the years, Astorino says it’s become increasingly difficult to work with.

“One of the biggest challenges is the scarcity of engineers who are both skilled in Scala and eager to work with it, which makes long-term maintenance a real concern,” he says. Rather than taking on the enormous risk and cost of rewriting everything at once, Auvik took a practical route: keeping the most stable, low-change parts of the system intact while steadily rewriting other sections in Go.

“This strategy allows us to balance reliability with modernization, ensuring business continuity while positioning our technology stack for the future,” he says.

But at ADAMnetworks, CEO David Redekop has taken a different approach to legacy debt. “Fortunately, we’ve been able to address technology debt except for some behind-the-scenes services relying on Python versions prior to 3.8,” he says. “The only reason these technologies are still in place is because we can still use them. They don’t represent an urgent business problem in that there’s no risk exposure.”

Redekop’s strategy involves restricting new deployments and updates in old code. That means anytime an update is imperative, that’s when code is refactored on an as-needed basis, he says.

The art of having it both ways

Smart CIOs know they don’t have to toss out reliable systems to get the benefits of new technology. Instead, they’re finding ways so they can work together, letting proven systems keep doing their jobs while modern tools handle the innovation.

Yahoo shows how this works in practice. “We’ve built a number of features in the cloud that integrate into how mail works,” Zen says. “So for example, TLDR, which summarizes emails for people, that’s an AI-based feature we built using cloud technologies, but without having to change the way our physical infrastructure works today.”

Over at IBM, CIO Matt Lyteson describes how the company has managed to make decades-old applications play nice with modern development practices. “We have many older applications operating on legacy systems integrated with a CI/CD pipeline, enabling secure and frequent code deployment into the environment,” he says. “Automation tools such as Ansible and Python are used to administer the core operating system and hardware.”

The integration strategies vary depending on what each organization needs. Brown-Forman focused on data integration, connecting its product lifecycle management system to a master data management platform that pulls information from multiple enterprise applications. This approach eliminated the manual data copying that used to eat up time and introduce errors, according to Homer.

Auvik demonstrates a sophisticated data synchronization approach. “We integrate modern capabilities by synchronizing all relevant data between the legacy system and our new platform in near real-time,” Astorino explains. “This ensures core operations remain stable within the legacy environment, while enabling centralized access to data for cloud-based analytics, AI-driven insights, and new product functionality. By decoupling the data layer from the legacy system, we avoid operational disruption.”

For ADAMnetworks, the integration philosophy centers on sustainability. The catch is you can always add and integrate new tools, Redekop says. It’s tempting because the quick wins are easy to grab, but you have to weigh both the short- and long-term costs, especially when those short-term gains come with the risk of building up tech debt, he adds.

“Our approach has been to balance new capabilities in a scalable and sustainable way with a clear path of methodically eliminating tech debt,” he says. “We often liken this to operating on the heart with the elevated importance of not killing the patient in the process.”

IT services company CDW has taken a similar approach with what senior VP and CTO Sanjay Sood calls the company’s system of engagement, essentially a flexible mid layer that lets it plug in new capabilities without reworking core systems.

“This architecture gives us the ability to decouple components, making it much easier to add or replace functions as our business evolves,” he says.

People make or break every tech project

When it comes to modernization, the most difficult challenges often aren’t technical, they’re human. Employees get comfortable with systems they’ve used for years, sometimes decades. Try to change too much too fast, and they’ll push back, disrupting even the best-laid plans.

Auvik’s Astorino has learned this lesson firsthand. “Culture and change management are pivotal when modernizing legacy infrastructure,” he notes. “Success depends on strong leadership alignment and the ability to guide teams through not just technical shifts, but mindset shifts as well.”

Astorino highlights a common pitfall. “A common challenge is people often expect the modern system to replicate 100% of legacy functionality, when in fact modernization should reflect today’s market needs rather than the assumptions baked into systems years ago.”

ADAMnetworks’ Redekop takes an even more direct approach to organizational alignment. “This is a huge and often misunderstood aspect of change in general,” he says. “Getting engagement and alignment with everyone that may be affected by a change can’t experience a shortcut. The effort of all rowing in the same direction for the same reason is a requirement that, when not met, will lead to bigger and worse problems than the one being solved.”

Helena Nimmo, CIO at IFS, is even more blunt. “It’s not just  changing technology,” she says. “It’s a behavioral change.” Her team invests heavily in stakeholder engagement, training, and building internal advocates who can help others navigate the transition.

CDW’s Sood has seen this dynamic play out repeatedly. His approach focuses on turning modernization into a growth opportunity that empowers teams to spend less time on maintenance, and more on innovative, value-added projects, he says.

Detective work: finding the real problem

One of the trickiest parts of working with legacy systems is figuring out whether problems are actually caused by the technology itself or by the data, processes, and people that have grown up around it over the years.

Astorino offers a practical diagnostic framework. “In my experience, challenges with legacy systems often stem from a mix of technology, data, processes, and people,” he says. “The key is to look at the symptoms: instability points to technology, inefficiency points to processes, blind spots point to data, and skill gaps point to people.”

Astorino breaks it down further. “It’s about technology when the system itself becomes unstable or increasingly hard to maintain,” he says. “It’s about people when critical expertise has left the organization and new staff struggle to operate or extend the system. Data becomes the issue when poor quality or rigid structures prevent insights and integration, and processes are the culprit when outdated workflows embedded in the system no longer align with how the business needs to operate today.”

ADAMnetworks’ Redekop takes a business-first perspective on system evaluation. “Charting the entire process from outset to outcome has to make business sense,” he says. “Ultimately, we’re always prioritizing eliminating biological bottlenecks, but it should always consider the whole, never just a small portion of the process.”

His philosophy centers on customer experience. “The right over-arching motivation ought to be the best business process possible for the organization’s objective,” he adds. “That objective must be dominated by the desire for fantastic client experience.”

Liberty Mutual has built what Palmer calls a proprietary tech debt management framework that evaluates multiple variables, including code quality, system age, business impact, data complexity, and ownership gaps. This systematic approach helps the insurance giant avoid the trap of assuming every pain point requires a technical solution.

Show me the money: when modernization pays off

An IT strategy only matters if it drives real business results. Companies that have modernized their legacy systems say the benefits are clear, with gains that flow straight to the bottom line.

Auvik shows how making the right changes can give customers real benefits. Astorino describes the company’s decision to modernize its alerting system. “A clear example of modernization delivering business value was our decision to move the alerting engine out of Scala and into our new Go services,” he says. “The architectural limitations of Scala made it nearly impossible to introduce meaningful enhancements, such as noise suppression, that our customers needed.”

The results were immediate and measurable. “By rebuilding the alerting engine in Go, we were able to introduce significant improvements, including a tolerance factor, time-in-state adjustments, and downstream alert suppression,” Astorino says. “These features not only reduced alert fatigue but also improved the accuracy and actionability of alerts, leading directly to better operational efficiency and customer satisfaction.”

ADAMnetworks demonstrates how modernization can enhance resilience and scalability at the same time. “One area where legacy systems carry clear value is in resilience and scalability, often addressed with the same modernization,” Redekop says.

The company’s approach includes implementing multi-zone, multi-cloud architectures with multiple service providers being load-balanced, he adds, offloading to a content delivery network, and making use of Anycast networks to deliver the fastest, best content that’s closest to the end-user.

Yahoo has seen dramatic improvements in development velocity, too. Teams that used to have to wait for hardware provisioning can now spin up cloud resources instantly. “We’re seeing faster shipping velocity across all of our business units,” Zen says. “We’ve definitely improved our time to innovation, speeding up how quickly we can build.”

Liberty Mutual’s contact center transformation shows how strategic integration delivers multiple benefits simultaneously. Instead of ripping out the company’s telephony platform entirely, the company modernized around it, achieving a double-digit percentage of calls now handled digitally, significant annual recurring savings projected through reduced infrastructure, lower maintenance, and fewer manual call-handling needs, Palmer says.

Over at Brown-Forman, they’ve achieved both cost savings and quality improvements through its data integration project, Homer says. By eliminating manual data copying between systems, they improved data quality while reducing the number of software licenses they needed.

The new rules of technology evolution

What organizations have figured out is whether you’re big or small, the approach to modernization is the same: decide what to keep, what to improve, and what to replace.

“As technology becomes more complex, it’s essential to take elements that make it more logical and simpler while simultaneously advancing,” says Redekop. “This may seem counterintuitive or even impossible, but it can be done. There’s majesty in simplicity and simplicity in majesty.”

This strategy requires patience, planning, and a solid understanding of the technology as well as the business needs. But the payoff is clear for organizations willing to make the leap to get the benefits of modern technology without the risks and costs of wholesale replacement.