By 2026, the enterprise software landscape across Asia-Pacific stands at an inflection point. The rise of agentic AI—autonomous systems capable of handling complex workflows from start to finish—is fundamentally reshaping how software is built, sold, and consumed.
Traditional per-seat licensing models are giving way to outcome-based, agent-orchestrated frameworks as organisations demand tangible returns from their AI investments.
The expectation gap driving change
The shift from seat-based to outcome-based pricing is not merely a commercial trend—it reflects a growing impatience with traditional software consumption models.
Simon Ma, regional head of APJ at Freshworks, points to a striking disparity revealed in the company’s May 2026 Global Cost of Complexity Report: 72% of mid-market executives expect clear ROI from AI within eight months, yet 55% admit that baseline deployment and integration take six to twelve months.
“Leaders cannot afford to wait, and they don’t want to pay for a human seat, a license where the AI agent is autonomously resolving, like, 80% of routine repetitive inquiries,” Ma explains. “The shift to demand of transparent pricing tied straight to business outcomes is actually here, and it is an active conversation that we have with every organisation today.”
This expectation gap is reshaping vendor strategies. Workday‘s APAC CTO Shan Moorthy observes a parallel evolution: customers are moving from fixed seat licences to consumption-based models where they “pay for AI based on the value it generates” rather than assuming agents run 24/7.
This pragmatic approach reflects a broader industry recognition that AI adoption must deliver measurable business value.
Architectural evolution for an agent-first world
The transition to agent-driven platforms requires fundamental changes to product architecture. For established vendors, the challenge lies in embedding autonomous agents without breaking existing enterprise stacks. Ma emphasises that Freshworks’ design philosophy prioritises interoperability, governance, and simplicity. “We cannot ask a company to tear up what they already have,” he states.
“They need to be able to snap directly onto existing systems and deliver fast value through embedding as much of the intelligence into their processes today.” Simon Ma
This pragmatic approach resonates across the industry. Moorthy describes a move away from monolithic architectures toward domain-specific platforms that work well within themselves and increasingly with each other through agentic layers.
“Traditionally, we would have done point-to-point integrations with APIs,” he notes. “But now we’re seeing the rise of AI agents as the integration point.”
Tackling data fragmentation
One of the most persistent challenges in AI adoption across Asia-Pacific is data fragmentation. Ma notes that the average mid-market company runs four to five different AI tools simultaneously, triggering what Freshworks terms the “AI complexity tax”—organisations lose approximately 25% of their AI budget globally to fragmentation, tool sprawl, and consultant overheads.
To address this, the industry is gravitating toward open, secure standards like Model Context Protocols (MCPs). “You can integrate frameworks like MCP gateways into all the different unified platforms, similarly to what Freshworks is doing with Freshservice,” Ma explains. “Then we can allow all these agents to securely connect to all these different data sources without much friction.”
This approach bypasses heavy custom deployment development, significantly lowers implementation timelines, and avoids unnecessary manpower costs—critical considerations in a region where labour shortages and cost pressures are driving digital transformation.
The security and sovereignty imperative
Data sovereignty and security concerns are reshaping innovation priorities. Ma references the concept of “Sovereign AI by Design,” arguing that true compliance is structural rather than merely a matter of server location.
“AI agents need deep operational context to be effective—we cannot leave them sitting in isolated point solutions. Naturally, they must be woven into core workflows and data layers.” Simon Ma
This requires building security from day one. Freshworks and other vendors are integrating audit trails, identity controls, and local compliance features directly into their platforms.
For highly regulated organisations in sectors like finance and logistics—including regional players such as Grab—these built-in controls enable confident automation adoption while maintaining governance.
The talent transformation
Perhaps the most significant shift involves the workforce. Ma argues that organisations must cultivate “AI orchestrators” from within their existing talent pools, leveraging employees with deep domain knowledge who can be trained to manage AI-driven workflows.
“Domain-specific knowledge is very important,” he emphasises. “For example, if an organisation is in the shipping or manufacturing industry, understanding that domain is crucial.”
Sea Limited‘s co-founder David Chen echoes this perspective, predicting a fundamental reconfiguring of engineering teams where developers evolve into “system orchestrators” who spend their time on product judgment, system design, and orchestrating AI-driven workflows rather than routine coding.
At Sea, which operates Shopee and other digital platforms across Southeast Asia, agentic AI coding tools are enabling developers to “think better, not just type faster”.
Measuring what matters
The shift to agent-first models demands new measurement frameworks. Ma advocates moving beyond traditional metrics like service-level agreements—which measure speed—toward “experience level agreements” that track three pillars: service reliability, service quality, and user effort. “Instead of just looking at how fast we resolve the ticket, we look at how well the ticket was solved,” he explains.
This evolution complicates ROI justification but reflects the deeper value AI agents can deliver. Organisations must track automation effectiveness—the percentage of workflows successfully resolved end-to-end by agents—and monitor whether AI deployment reduces fragmentation rather than adding to the 25% budget loss already observed globally.
Looking ahead
As Asia-Pacific enterprises navigate this transformation, the best-positioned organisations will choose platforms that are powerful yet easy to administer, anchored around unified architectures. Ma emphasises the importance of getting fast wins and learning along the journey.
“Organisations that adopt an AI-first culture naturally have an advantage,” he concludes. “As we have seen time and again, organisations that adopt new things and stay agile are more resilient and grow better in the long term.”
The agent-first revolution is not merely about technology—it represents a fundamental shift in how software creates value, demands new governance frameworks, and ultimately promises to decouple business growth from linear headcount growth.
