Who's AI ready? Apparently just 13% of firms globally, says study
The latest Cisco AI readiness report highlights a divide between companies' growing AI ambitions and a lack in infrastructure needed to make them a reality.
AI Artificial Intelligence words are seen in this illustration taken May 4, 2023. (Photo Illustration: Reuters/Dado Ruvic)
This audio is generated by an AI tool.
SINGAPORE: Despite rapid adoption of artificial intelligence across industries, most companies around the world are still struggling to turn their AI ambitions into real business value, a new study shows.
Cisco's AI Readiness Index 2025, released this week, found that just 13 per cent of organisations surveyed are fully prepared for the AI age – meaning they have scaled AI beyond pilot projects and achieved measurable results.
The remaining firms are still experimenting with limited outcomes or lagging behind due to gaps in data infrastructure, security and governance.
Overall, Southeast Asia is performing slightly better than the global average, with about 16 per cent of companies in the region classified as AI-ready.
The tech giant’s latest report – into its third edition – surveyed more than 8,000 technology and business leaders across 26 industries and 30 economies.
“Clearly, we're not as ready as we'd like to be,” said Ben Dawson, Cisco’s president for Asia Pacific, Japan and Greater China.
“(Companies) with the highest state of readiness … are looking at AI not as an adjunct to their business, but as a core part of the business strategy,” he added.
“(They’re) investing appropriately, having the right policies and governance … and getting security embedded into their infrastructure.”
SOUTHEAST ASIA IN THE LEAD
While the study assessed businesses rather than countries, Dawson said some economies are laying stronger foundations for AI adoption.
For instance, Indonesia and Thailand – with AI readiness levels of 23 per cent and 21 per cent respectively – outperformed their European and American counterparts, which stood at 11 per cent and 14 per cent.
“That says to me there are organisations in developing markets that see this as an opportunity to leapfrog, perhaps more than their developed peers,” he told CNA’s Asia First.
“(This) would also indicate that some of these markets are not dealing with the same level of tech debt that we're seeing in some larger organisations.”
Technical debt refers to the future cost of extra time and effort required to fix or rebuild systems due to earlier shortcuts or suboptimal solutions.
DEMAND EXCEEDS INFRASTRUCTURE
The report warned that AI infrastructure debt is becoming a growing risk, with bottlenecks including limited graphics processing unit (GPU) capacity, fragmented data systems and networks unable to scale fast enough to meet demand.
About 83 per cent of organisations surveyed plan to deploy AI agents, and almost 40 per cent expect these systems to support human workers within the next year.
Yet, only 34 per cent feel their company’s technology infrastructure is fully adaptable and scalable to accommodate the evolving computational needs of such projects.
Training is another key challenge.
Only one in three firms have a formal change management plan to guide employees through AI adoption – a gap that could hinder efforts to integrate AI into daily workflows and deliver real-world results.
BETTER GOVERNANCE NEEDED
Cisco noted that companies with strong digital foundations and well-defined governance models are the ones realising the most tangible returns from AI.
These organisations are also moving quickly from proof-of-concept trials to production, while others remain stuck in pilot phases.
Such businesses are capable of measuring and tracking the impact of their AI implementation, and are more confident of monetising those initiatives.
Dawson emphasised the need for greater collaboration across sectors to build responsible and resilient AI ecosystems so they can be adopted more widely.
“Embracing AI fully is definitely going to require a collegiate approach between technology companies, industry and government,” he said.
He added that effective regulation and cybersecurity frameworks are essential to building trust and enabling safe AI growth.
“Organisations … that do it well are looking at cybersecurity as an inherent, foundational element of that infrastructure,” he said.
“That requires getting the right policies in place from the government, getting the right level of investment (and) getting the right technology.”
