Before Indonesian enterprises can capture the returns from AI investment, the integration infrastructure beneath those initiatives needs to be addressed.
Indonesia published its AI National Roadmap White Paper in July 2025. Microsoft committed USD 1.7 billion to cloud and AI infrastructure in-country, the company’s largest investment in Indonesia’s history. NVIDIA and Cisco co-funded an AI Centre of Excellence with Indosat, targeting AI access for hundreds of millions of Indonesians by 2027. The government’s investment case is unambiguous. Digital transformation spending is forecast to reach USD 29 billion in 2026, growing to USD 69.57 billion by 2031.
However, there is a large constraint that will determine how much of that investment translates into actual enterprise outcomes: most Indonesian organisations are still running integration infrastructure designed for a different era. The AI tools are ready, and the talent and capital are moving. The stack underneath is the problem.
What the AI Era Requires from Integration
When organisations talk about AI transformation, the conversation focuses on which model to use, what use cases to prioritise, and how to govern the outputs. What gets less attention is the plumbing: the integration layer that connects AI agents to the systems they need to act on.
An AI agent doesn’t produce value by answering questions, but by taking actions: pulling inventory data from a legacy ERP, updating a customer record in a CRM, triggering a compliance check in a banking system, and routing a decision to a human when it needs one. Each of those actions requires a reliable, governed connection between the agent and the underlying system. That’s what integration infrastructure does, and it’s what legacy platforms were not built to handle at the speed and flexibility that AI requires.
Indonesia’s financial sector provides a useful frame. The OJK’s digital banking mandates and Indonesia’s Personal Data Protection Law — which came into full effect in 2024 — require that enterprises maintain clear audit trails and data governance across integrated systems. Running AI agents on top of legacy ESB architectures, where governance is centralized and brittle, makes compliance harder, not easier. Integration modernisation doesn’t just enable AI, but also builds the foundation for regulatory compliance.
IDC’s FutureScape 2026 analysis puts the stakes directly: by 2030, 50% of new economic value generated by digital businesses in Asia-Pacific will come from organisations investing in and scaling AI capabilities today. By 2027, IDC projects that 50% of Asia-based CIOs will be tasked with creating enterprise AI value playbooks, not just pilot programmes, but executed strategies that show measurable business impact. Enterprises that still cannot connect AI to business systems will lag behind.
The Legacy Stack Problem in Indonesian Enterprises
The integration challenge looks different in Indonesia than in more mature markets. Many of the country’s largest enterprises built their IT infrastructure over multiple generations of middleware. Some are running TIBCO or webMethods (now part of IBM following Software AG’s 2024 divestiture) at their core. Others are on first-generation cloud iPaaS platforms, which were built before AI orchestration was a consideration.
Workato’s analysis of what’s holding enterprises back on legacy ESB platforms identifies several patterns that are directly relevant to the Indonesian context. Legacy platforms require deep, proprietary expertise to build and maintain. That expertise is expensive and scarce in most markets, and the scarcity is real in Indonesia, where demand for senior integration and DevOps engineers pushed salaries into double-digit growth in 2025. Organisations relying on a small group of specialists to keep their integration layer running are one talent departure away from a critical dependency problem.
Deployment timelines compound this. New integrations on legacy platforms routinely take 12 to 18 months from requirements to production. For a retailer trying to connect Tokopedia or Shopee data to its ERP for real-time inventory decisions, or a bank needing to connect a new digital lending product to its core banking system within weeks of regulatory approval, that timeline isn’t a technical inconvenience. It’s a competitive disadvantage.
Connectivity as the Foundation, Not the Ceiling
The framing that matters in the Indonesian context is this: automation and AI are accelerators of growth, not replacements for human capability. The organisations getting this right are using modern integration platforms to expand what their teams can accomplish: connecting systems that were previously siloed, surfacing data that was previously inaccessible, and giving people the information they need to make better decisions faster.
Gartner projects that by 2030, 70% of enterprises will pivot to a consolidated automation platform that orchestrates AI agents, business processes, and APIs in a single environment, up from roughly 5% today. The enterprises moving first are finding that the consolidation investment delivers value in two places simultaneously: it reduces operational overhead from maintaining fragmented stacks, and it creates the architectural readiness to deploy AI without rebuilding the foundation each time.
Avid Technology, a US-based media software company, provides a concrete reference for what this transition looks like in practice. When vendor support for their webMethods ESB was ending, they attempted a migration to another enterprise platform that was never fully implemented due to complexity and specialist dependency requirements. They then moved to Workato. Their CIO described the outcome: “Now we have a greater capacity to support digital business growth without worrying about scalability and performance of middleware.” The migration that looked like a disruption became the infrastructure that enabled growth.
The velocity difference is consistent across organisations making this transition. Workato customers report integration development cycles that run four to ten times faster than the legacy platforms they came from, with fewer specialist dependencies and significantly less maintenance overhead. A director of enterprise technology at a 5,000-person company described building 70-plus integrations across eight systems in eight months after consolidating onto a single modern iPaaS, with 20% cost savings from eliminating redundant tools.
The Integration Question Comes Before the AI Question
Gartner’s data on AI agent adoption puts a deadline on the conversation: 40% of enterprise applications will be integrated with task-specific AI agents by end of 2026, up from under 5% in 2025. Separately, Gartner projects that more than 40% of agentic AI projects will be cancelled by the end of 2027, with inadequate risk controls and unclear business value as the primary drivers. The organisations that succeed will be the ones that built their integration layer before they needed it for AI, not the ones trying to retrofit governance and connectivity onto fragile middleware after the fact.
Indonesia’s enterprise IT environment is entering what the Digital Transformation Summit Indonesia describes as a pivotal phase: shaped by the PDP Law compliance mandate, rapid AI adoption, and the government’s Golden Indonesia 2045 vision. The organisations making smart platform decisions now will be the ones able to run AI at scale three years from now.
Workato’s platform is built for this transition: a unified orchestration environment that handles integration, automation, and AI agent coordination across cloud and on-premise systems, with governance and auditability built in. For Indonesian enterprises still running legacy middleware, the three-part whitepaper series on moving beyond ESB platforms covers the risk assessment, migration approach, and what to look for in a platform that can carry the weight of both today’s operations and tomorrow’s AI initiatives.
The investment case for AI in Indonesia is settled. The integration foundation that makes it work is where the real decisions are being made.