In the bustling corridors of a recent ServiceNow roundtable in Malaysia, a clear consensus emerged among the nation’s top CHROs, CFOs, and CIOs: the era of tentative AI experimentation is over.
The conversation has pivoted decisively from “Can we use AI?” to “How do we become an agentic business?”
This shift goes far beyond merely adopting new technology; it demands a fundamental reimagining of talent, governance, and the very workflows that define an enterprise.
Tahsin Alam, head of ASEAN for ServiceNow, opened the dialogue by challenging a common misperception: that Asia lags in digital innovation. “I think everybody thinks that just because we’re in Asia, we’re lagging,” he noted, pointing to the “really interesting stuff” happening in the region.
He observes that customers are modernising IT systems and moving beyond traditional ticketing to transform HR services, creating streamlined, end-to-end processes where employees can get answers anytime, aided by a person or an AI agent.
This narrative is central to the theme of the roundtable, “The AI Control Tower for Business Reinvention: From AI Experiments to an Agentic Business.” The goal isn’t to replace humans but to elevate talent into higher-impact roles focused on strategy and innovation, using AI as a “strategic multiplier.”
The examples from ServiceNow’s own operations are compelling. As Alam shared, their finance purchase order process was slashed from two weeks to just a few days, and HR query resolution became five times faster.
The key? They didn’t just bolt AI onto an existing process. They “essentially looked at the whole workflow logic and built it from the ground up”. This distinctionārethinking work from the coreāis what separates true reinvention from superficial enhancement.
The blueprint for an agentic business: A unified platform
The urgency for this reinvention is underscored by the challenges enterprises face. The 2026 ServiceNow Enterprise AI Maturity Index reveals that while enterprise AI investments are projected to top US$2.5 trillion, only a small fraction of organisations have effectively replaced their legacy systems.
The primary obstacle? A “patchwork enterprise” where hundreds of apps create a fragmented, ungoverned landscape.
To move from promising prototypes to production-ready transformation, a new architectural blueprint is required: one that is sense, decide, act, and secure. At its heart is the principle that “AI without workflows is just expensive advice. AI inside workflows is autonomous enterprise execution”.
This requires a single, unified platform that connects data, workflows, and governance across the entire enterprise, not a series of siloed, bolted-on solutions. This is the essence of the ServiceNow AI Platform, designed to be the “AI control tower for business reinvention“.
Governance: The trust layer for autonomous work
As discussions at the roundtable moved toward the “autonomous enterprise,” the conversation naturally turned to trust, accountability, and control. Alam defined autonomy in this context: “autonomous essentially means that an AI agent is completing the end-to-end tasks with defined rules, defined permissions, and obviously guardrails”. It is no longer just a simple assistant.
However, a significant gap exists between employee adoption and enterprise readiness. “The resounding answer to that in this session was, most certainly not,” Alam stated. “The constraint wasn’t the access to tools… it really is around governance clarity”.
Employees are hesitant, with research cited showing 48% feel uncomfortable admitting they use AI at work for fear of being seen as lazy or incompetent. This is a governance and culture issue, not a skills gap.
To address this, the discussion highlighted a new class of AI specialists designed to execute complete workflows across IT, CRM, employee services, and security. These aren’t just task-level helpers; they take on whole sets of workflows. This advancement puts a premium on governance.
The need for a “control tower”
When multiple AI agents from ServiceNow, Salesforce, SAP, and others proliferate across an enterprise, the potential for chaos is immense. Alam emphasised the necessity of a central governance layer: an AI control tower.
“Now that you have all these proliferations of agents or AI that have been built by many systems within your organisation, what happens if something goes wrong? Can you reverse it? Can you tell your board… that this agent acted within this defined scope?” Alam asked.
This is where the concept of the AI Control Tower becomes critical. It provides the ability to discover AI assets across any cloud or enterprise system, observe their behaviour, enforce security policies, and even block rogue agents.
The recent partnership with NVIDIA to integrate NVIDIA OpenShell as a secure runtime is a direct response to this need, creating a “trust layer” for autonomous AI. As one observer noted, “AI agents need the platform more than humans do… Without governance built into the platform, a brilliant agent becomes a brilliant liability”.
Similarly, industry analysts are also supporting the idea. Forrester describes an Agent Control Plane, which is a centralised layer that governs, monitors, and orchestrates AI agents across enterprise systems.
“As agents proliferate across the “build” plane and the “orchestration” plane, governance must sit outside both planes in order to provide independent visibility, enforce consistent policies, and maintain control when runtime environments behave unpredictably. This creates a distinct need for an out-of-band “oversight” plane that can enforce policy and maintain trust ā regardless of how or where agents are built and executed,” shared Leslie Joseph, principal analyst at Forrester.
A tale of transformation: The NTT DOCOMO and StarHub case
A powerful real-world example of this blueprint in action is the recent collaboration between NTT DOCOMO, StarHub, and ServiceNow. This initiative tackles the age-old frustration of international roaming. When a customer loses service overseas, multiple carriers with different systems and web forms must coordinate to fix it, leading to delays and poor customer experience. This is a classic example of fragmented, inter-enterprise workflows.
The three companies developed a shared operational model on the ServiceNow AI Platform. They automated and standardised the process for resolving roaming faults, turning manual coordination into an autonomous, real-time workflow.
When an issue occurs, the system quickly shows what happened, which network is affected, and what is being done. This proactive approach benefits everyone: travellers get reliable connectivity, and operators gain efficiency and improve customer trust. The companies targeted a commercial launch for the second half of 2026, creating a scalable model for global interoperability.
The path forward: A shared and unified vision
For Malaysian leaders, the lessons from the roundtable and these examples are clear. The journey to an agentic business is not about a single AI project. It is about adopting a comprehensive platform that allows for a crawl, walk, run approach, where the platform and governance are in place from the start.
The ownership of this transformation must be shared. Alam stressed that accountability must be predefined, not assigned after the fact, and must go “all the way from the top, from the board level to the CIOs, CFOs, because no single function can own this alone”.
The organisations winning on AI are those solving fundamental architecture and governance questions first. With the proper foundationsāa unified platform, robust governance, and a culture that empowers employeesāMalaysian enterprises can confidently navigate the shift from AI experimentation to becoming truly autonomous, agentic businesses capable of reinvention.
