Malaysia’s telco sector has gone through various structural changes in the last five years where it has seen its top three wireless telcos merge and the shift from a 5G single wholesale network model to a dual network model.

With the merger, the Malaysian telco landscape has become increasingly competitive. Postpaid average revenue per user and the number of prepaid customers have both declined, while competition in fibre-to-the-home services has ramped up as telcos roll out their fixed mobile convergence strategy.

Furthermore, the move to the dual network model means that telco operators on Digital Nasional Berhad’s 5G network will need to continue to ensure optimal network experience to stay competitive while U Mobile will need to balance demand and avoid network congestion as they grow.

The challenges in Malaysia’s market are also echoed by global trends. Deloitte’s 2026 Technology, Media and Telecommunications Predictions report found that up to 77% of global consumers feel no real loyalty to their provider, only 47% stay with the same operator for more than five years, and annual churn across the industry sits at around 22%. At the same time, 86% of consumers say they are willing to pay more for a better experience.

These dynamics have placed strong pressure on telcos to innovate and improve their performance. In response, agentic AI offers possible solutions across the value chain.

Autonomous AI agents can enhance sales with product recommendations hyper-personalised to individual customers, streamline operations by managing network traffic and issues more autonomously to improve network quality and reduce downtimes. Unlike traditional AI tools that support individual tasks, agentic systems can coordinate actions, execute workflows, and resolve issues across functions end-to-end. This shifts AI from a back-office capability into a core operational engine – one that can fundamentally change how telecom organisations work.

The stakes are significant. Industry estimates suggest the global market for autonomous AI agents could reach US$8.5 billion (RM34.7 billion) in 2026 and US$35 billion by 2030, with Deloitte projecting that figure could climb as high as US$45 billion for organisations that orchestrate agents well. The value sits not with those that experiment most, but with those that move fastest from experiment to production.

Globally, many telecom operators are stuck in “pilot purgatory” – running multiple AI initiatives that show promise individually, but fail to scale or deliver meaningful business outcomes. The issue is not innovation - it is execution discipline: funding is fragmented, ownership is unclear, and pilots that show promise never reach production.

However, there are some operators making headway, by taking a different approach. They are treating AI as a business capability, not a technology experiment. They come up with a clear, organisation-wide AI roadmap with integrated workflows, and stick to it. They assign clear ownership, align initiatives to measurable outcomes, and make disciplined decisions about which use cases to scale and which to stop.

In Malaysia, we have observed telcos taking various approaches in adopting AI. One has partnered with hyperscalers to release an AI sandbox for its employees to experiment, innovate, and implement generative AI solutions; while another has signed memoranda of understanding with network vendors to use AI core features to improve their network stability efficiency and stability.

While there is no one-size-fits-all answer, telcos should keep in mind three key ingredients as they continue on their AI journey.

First, they need to find the right partner ecosystem rather than isolated tools. Second, initiatives should be designed for production from day one, not open-ended experiments. Third, they need teams empowered to innovate within clear guardrails on cost, risk, and compliance.

The window for action is relatively short. Over the next 12 to 18 months, operators who successfully embed these capabilities will build advantages that competitors will struggle to replicate. Those who delay will fall behind – not because they lack access to technology, but because they lack the organisational discipline to use it effectively.

Piyush Jain is the technology, media and telecommunications industry leader for Deloitte Southeast Asia’s Strategy, Risk and Transactions business. He also leads the Monitor Deloitte practice in Southeast Asia and is an experienced adviser to CxOs in the region. Nicholas Ong is a technology, media and telecommunications industry partner in Deloitte Malaysia’s Technology and Transformation business. Aditya Shankar is a director at the Monitor Deloitte practice in Southeast Asia, where he advises leading banks and telecom operators across Asia-Pacific on AI strategy, value realisation and responsible deployment.