Malaysia’s digital economy is growing fast and AI investment is already committed. The integration layer underneath those initiatives will determine how much of that spending becomes measurable business impact.
Malaysia’s digital transformation market is projected to grow from USD 12.67 billion in 2026 to USD 29.74 billion by 2031, and AI sits at the centre of where that investment is heading. The enterprise conversation has already shifted from whether to adopt AI to how to make it work at scale. The answer to that question runs through a part of the stack that rarely leads the planning discussion: the integration layer.
Getting this right is what converts AI investment into business outcomes. A modern integration platform gives agents real-time access to the enterprise systems they need to act on and keeps every action governed and auditable, without the specialist bottlenecks that slow legacy stacks. The enterprises building that foundation now will deploy AI at a pace that older platforms simply cannot support.
The Platform That Got You Here
Platforms like TIBCO, webMethods, and older generations of enterprise service buses were built for a specific context: centralized IT teams managing every connection, integrations built by specialists, deployments measured in quarters rather than weeks. In that world, they worked well. They brought structure to complex on-premise environments and ran reliably for years.
The issue isn’t reliability. The issue is that AI agents operate on a fundamentally different pattern. When an agent needs to pull customer data from Salesforce, cross-reference it against an SAP ERP record, trigger a workflow in ServiceNow, and log the outcome for audit, in a single, real-time sequence, a legacy ESB introduces latency, fragility, and governance gaps that make the whole thing impractical at scale. The middleware was never designed for that kind of multi-system, event-driven coordination.
Gartner’s research on the integration landscape puts a number on where this is headed: by 2030, 70% of enterprises will pivot to a consolidated automation platform that orchestrates AI agents, APIs, and business processes together, up from roughly 5% today. The organisations moving first are discovering that the integration layer isn’t just a technical decision; it’s the architectural foundation that determines how quickly AI can be deployed and how reliably it performs in production.
What Staying on Legacy Actually Costs
The case for staying on a legacy ESB usually comes down to three things: the existing integrations work, the migration seems risky, and there’s always something more pressing to tackle. Those arguments made sense when the cost of inaction was slower delivery timelines. They look different when the cost is AI initiatives that never move past proof of concept.
Workato’s analysis of legacy ESB environments identifies several patterns that emerge consistently across organisations running TIBCO, webMethods, or similar platforms. Deployments routinely stretch to 12 to 18 months, not because the integrations are complex, but because the platform requires specialist expertise and centralised governance that slows every step. New team members can’t build independently. Minor upgrades require full regression testing. And the total cost of ownership, including licences, consultant dependency, maintenance overhead, keeps climbing even as the platform’s strategic value to the business declines.
Vendor changes compound this. When Software AG sold webMethods to IBM, customers got new commercial terms, a different roadmap, and uncertainty about long-term product direction. That’s a familiar pattern in this market. Salesforce’s acquisition of Informatica creates similar questions for organisations in the MuleSoft ecosystem: what happens to roadmap priorities, licensing structures, and the independence of integration capabilities when a platform becomes part of a broader CRM-focused portfolio?
The answer, for most organisations on the receiving end of those consolidations, is that the migration they put off becomes unavoidable and considerably more expensive for having waited.
The Migration Doesn’t Have to Be a Rip-and-Replace
Atlassian, the Australian enterprise software company behind Jira and Confluence, ran into this constraint directly when it needed to migrate its core ERP from NetSuite to Oracle Fusion Cloud. The project had a nine-month window and an integration landscape carrying over 800 business process changes. Their previous iPaaS had accumulated enough specialist dependency, slow development cycles, and technical debt that it couldn’t carry the migration at that pace. Moving to Workato, the team completed 50 integrations within the nine-month window without disrupting financial operations. Claude Khoury, Atlassian’s Head of Finance Ops Enablement & AI, described the cumulative impact: “Workato has been a key part of our transformation journey, helping teams across the business save over 100,000 manual hours and deliver integrations and automations faster than we thought possible.”
The organisations doing this well aren’t treating stack modernisation as a discrete project with a cutover date. They’re running a modern integration platform alongside existing infrastructure, proving value on new workloads, and migrating integrations as contracts renew and systems are upgraded. The risk profile is manageable when the approach is incremental.
A director of enterprise technology at a 5,000-person company described what consolidating onto a single iPaaS delivered after migrating from MuleSoft: seventy-plus integrations across eight systems, built in eight months, delivering 20% cost savings through reduced tool sprawl and eliminated redundant licences. Development velocity was the other number that changed: Workato customers consistently report integration cycles that run four to ten times faster than the platforms they came from.
The AI Readiness Question
Gartner projects that 40% of enterprise applications will be integrated with task-specific AI agents by end of 2026, up from under 5% in 2025. For Malaysian IT leaders managing the country’s manufacturing base, financial services stack, and regional business operations, the timeline is concrete. The AI initiatives that boards are asking about now will require integration infrastructure that most legacy platforms aren’t designed to provide.
Workato’s research on why integration platforms become the foundation for agentic AI makes the dependency chain clear: the organisations moving fastest on AI aren’t chasing the newest agent frameworks. They’re building on integration foundations they already trust and layering AI deliberately, with governance built in from the start. The consolidation investment pays back not just in reduced operational overhead, but in the architectural readiness to scale AI without rebuilding the foundation every time a new initiative starts.
For Malaysian CIOs still running TIBCO, webMethods, or first-generation iPaaS, the relevant question isn’t whether to modernise. The question is how to sequence the transition so that AI initiatives can actually start moving. Workato’s three-part whitepaper series on moving beyond legacy ESBs covers the risk assessment, practical migration path, and what to look for in a modern orchestration platform.
The window for making that transition ahead of AI commitments rather than because of them is narrowing.