AI agents vs automation
Uncover the key differences between AI agents and automation. Learn how each technology can improve workflows and drive smarter decisions for Australian businesses.
Summary: As generative AI moves from experimentation to enterprise-wide deployment, Australian IT leaders are realising that legacy infrastructure and fragmented data are holding them back.
This article explores why successful GenAI adoption in the years ahead hinges on modernising cloud environments, unifying data platforms, and embedding governance at every layer. Drawing on expert insights from Canon Business Services ANZ, it highlights real-world examples, common misconceptions, and the critical role of MSP partnerships in accelerating value.
The message is clear: without the right foundations, AI can’t scale, but with the right strategy, modernisation becomes a launchpad for transformation.
From boardrooms to backends, Australian organisations are facing a pivotal moment: the hype around generative AI has given way to serious investment and serious challenges.
Yet as CIOs and CTOs rush to experiment with new tools and large language models, many are discovering a hard truth: you can’t innovate your way out of outdated legacy systems and infrastructure.
Today, it’s no longer enough to migrate to the cloud or dabble in data lakes. To compete, IT leaders must build for a future defined by scalable, secure, AI-powered platforms.
And that means confronting legacy debt, messy data estates, and cultural hesitation, before Gen AI can drive real value.
While the appetite for Gen AI is high, the foundations are often missing. According to Rico Gunawan, Head of Modern Applications at Canon Business Services ANZ, mindset, not new technology, is the biggest blocker.
“Many organisations still treat GenAI as a shiny add-on rather than a strategic enabler,” he says.
“They underestimate the foundational work required: clean, accessible data, robust governance, and a clear use case. Without these, GenAI becomes a proof-of-concept treadmill.”
CBS Head of Technology Solutions, Adrian Capolino agrees. In his view, it’s not a lack of enthusiasm holding development teams back. It’s that expectations outpace readiness.
“They jump in with pilot projects and big hopes, but the data’s messy, the infrastructure’s outdated, or the development teams don’t have the skills. Things stall because the foundations just aren’t there.”
And even in organisations further along the journey, hybrid environments and entrenched legacy systems are complicating progress.
“Most IT leaders are still managing hybrid environments, some workloads in the cloud, others duct-taped together,” says Rico.
“Being AI-ready means having the courage to retire what no longer serves the business. That’s not easy when legacy platforms are deeply embedded.”
Cloud is no longer a “destination.” It’s the operating model. But GenAI takes that a step further. As Adrian explains, getting AI-ready is about rethinking how the tech stack supports continuous innovation.
“GenAI demands scalable compute, unified data access, and modern architecture. But many businesses are still working with siloed data, rigid processes, and complex legacy systems that weren’t built for this kind of load.”
Rico frames it as a balancing act: modernising with an eye to the future while still maintaining what keeps the business running today.
This means legacy modernisation strategies need to be both practical and intentional, not tech-for-tech’s-sake.
“It’s not just about throwing more compute at the problem,” he says. “You need integration, orchestration, and governance. If the platform isn’t ready, GenAI just exposes those cracks.”
It’s easy to assume that running GenAI in the cloud is as simple as switching on GPUs. But Adrian cautions against that mindset:
“Just having cloud access doesn’t mean you’re ready to run GenAI at scale. You need fast networking, scalable storage, clean data, and clear governance. Licensing costs can spike quickly if workloads aren’t optimised, and not all platforms are created equal when it comes to AI capabilities.”
Being in the cloud isn’t the same as being cloud-optimised for AI workloads. The right cloud strategy needs to include:
For Rico, the era of the “big bang” is mostly behind us.
“Most clients are phasing modernisation, starting with high-impact, low-risk apps to build momentum,” he says. “We’re seeing continuous modernisation becoming the norm, where it’s treated as an ongoing capability rather than a project with an end date.”
This iterative, outcomes-led approach enables teams to learn as they go, integrate new capabilities faster, and justify investment based on real business needs and value, not just infrastructure milestones.
One compelling example comes from CBS’s work with the South Australian Motor Sport Board for the bp Adelaide Grand Final, (formerly VAILO Adelaide 500).
Previously, the event relied on manual processes, Excel spreadsheets, and sticker-based credentials to manage access for contractors, competitors, and officials. The system was error-prone, time-consuming, and impossible to scale.
CBS delivered a fully digital solution using Microsoft Power Platform integrating Power Apps, Power Pages, Power Automate, and Dynamics 365.
“We didn’t just digitise forms. We reimagined the entire operational model,” says Rico.
The result was a scalable, secure platform that processed tens of thousands of credentials during the event. With integrated CI/CD pipelines, real-time data flows, and AI-ready architecture, the system is now poised to support AI-driven analytics and automation well beyond event operations.
Behind every AI initiative is data. And without strong governance, that data becomes a liability.
“Data governance isn’t just about organisation. It’s your first line of defence,” says Daniel D’Souza, Head of Information Security Solutions at CBS. “If you don’t know where your data comes from or how it’s being accessed, you’re creating blind spots. That’s creating security vulnerabilities and compliance risk, especially when AI is involved.”
So where does a strategic Managed Service Provider (MSP) fit in?
According to Adrian and Rico, the value lies in speed, specialisation, and de-risking.
“An MSP brings deep expertise across cloud, data, and security, and proven frameworks that reduce trial and error,” says Adrian. “We also help manage governance and compliance from the start, so internal teams can stay focused on business requirements and priorities.”
Rico adds that external perspective is often what clients need most:
“We’ve seen what works across industries. That lets us shortcut the guesswork and focus on impact. We also bring access to funding programs and accelerators many organisations don’t even know exist.”
Modernisation is no longer just about cost efficiency or system upgrades. It’s the foundation for innovation. And looking ahead, innovation increasingly means leveraging Gen AI that’s embedded, integrated, and intelligent, not bolted on.
But to get there, IT leaders must first address the fundamentals: clean data, modern platforms, robust governance, and cloud-native elasticity.
As Adrian puts it, “You can’t fake your way to artificial intelligence maturity. The foundations have to be real.”
With the right strategy and the right partner, those foundations can enable more than transformation. They can unlock competitive advantage that scales.
Legacy systems often contain decades of intertwined legacy code, business logic, and hidden technical dependencies that make change complex. These systems may not integrate easily with modern cloud architectures or AI workloads, creating bottlenecks in the modernisation process. Over time, they accumulate technical debt, limiting scalability, performance, and data security, all critical to running generative AI and modern analytics at scale.
When used effectively, generative AI (GenAI) can play a significant role in accelerating legacy modernisation. AI models can analyse legacy code, identify redundant patterns, and even generate clean, maintainable new code based on established business rules. By leveraging GenAI, organisations can automate documentation, highlight optimisation opportunities, and reduce manual effort, helping modernisation teams move faster while ensuring better code quality and consistency.
Replacing or refactoring legacy code allows teams to build modernised applications that are easier to maintain and integrate with existing systems. Frameworks such as Spring Boot code and domain‑driven design approaches can enhance scalability and performance while improving developer productivity. This also reduces the technical debt that slows innovation and enables businesses to stay competitive in a rapidly evolving digital landscape.
Successful modernisation efforts depend on balancing stability with innovation. While maintaining existing infrastructure for continuity, IT leaders progressively modernise workloads using cloud‑native tools, AI services, and containerised applications. This phased approach helps manage costs and ensures modernisation teams meet ongoing business needs without disrupting mission‑critical operations—turning the modernisation journey into a sustainable, continuous improvement process.
The first step in any legacy modernisation journey is to assess existing code, architecture, and business logic to determine what should be refactored, re‑platformed, or retired. Involving subject matter experts early helps uncover hidden technical dependencies and ensure informed decisions about future design. From there, define a clear target architecture, outline the scope of mainframe modernization if needed, and embed strong governance to reduce costs and improve accuracy along the way.
