Tata Communications and Bloomberg Media Studios have published a study on barriers to scaling artificial intelligence in large enterprises. The research found that 77% of senior executives now treat AI as a board-level priority.
The study surveyed 501 senior executives in Asia, North America and Europe involved in telecom infrastructure and procurement decisions at companies with revenue above USD $500 million. The findings point to a mismatch between AI spending plans and the systems needed to support them: 65% of respondents still operate on legacy or developing infrastructure.
Only 29% said their infrastructure can scale with changing business demands. The report argues that this is a major obstacle for companies trying to move beyond pilot projects, as AI workloads can grow quickly and expose weaknesses in connectivity, data architecture and system integration.
The research identifies five areas that determine whether AI investment produces lasting returns: infrastructure modernisation, interoperability across systems, skills, governance and visibility of return on investment. Progress in one area, it notes, can be undermined if another remains weak.
Legacy strain
Infrastructure modernisation emerged as a central fault line. Fewer than half of respondents said their organisations had fully modernised network connectivity, hybrid deployment flexibility or data architecture.
The study also found a clear divide between companies with newer systems and those still reliant on older technology. Enterprises with advanced infrastructure were nearly twice as likely to report high business value from AI as those operating on legacy systems.
Integration was another recurring problem. Some 28% of executives said difficulty integrating AI with legacy systems was a primary roadblock to creating value, while 38% said integration concerns contributed to delays in approval and procurement cycles.
Two-thirds of respondents, or 67%, said the seamless blending of digital automation and human interaction across channels is critical to AI execution. This suggests companies are struggling not only with technical deployment, but also with fitting AI into existing customer and employee processes.
Governance delays
The study found that governance structures are slowing decision-making as AI moves up corporate agendas. While board attention has increased, approval processes are becoming longer and more complex.
Security and compliance reviews were identified by 42% of respondents as the largest source of approval delays. Integration concerns and procurement complexity were each cited by 38%, suggesting the bottleneck often lies in internal processes rather than in a lack of interest in AI investment.
Skills shortages add another layer of pressure. Overall, 30% of enterprises said skill gaps and a shortage of specialised talent were a main barrier to realising AI value.
The challenge was more pronounced among larger companies. Among enterprises with revenue above USD $5 billion, 45% cited the skills gap, suggesting that scale itself increases the difficulty of finding and deploying specialist staff.
Value questions
Most companies reported at least some benefit from modernisation efforts, but many are still unable to track where those benefits are being created. Nine in 10 organisations said they see some value from modernisation initiatives, yet more than six in 10 said they have not reached optimal outcomes.
This is partly because AI, infrastructure and security are often measured separately. When those programmes are tracked in isolation, executives can struggle to see the broader effect across the business and may direct reinvestment based on incomplete signals.
For companies in Asia, the findings come as governments and businesses increase their focus on AI governance and regulation. The study included respondents from Singapore, China, Hong Kong and India, alongside executives in North America and Europe.
Singapore is among the markets where these tensions are becoming more visible as AI policy activity increases. This includes work on an AI agents registry for public officers, the establishment of the Singapore Artificial Intelligence Association and wider regional discussions linked to the ASEAN digital economy agenda.
Sumeet Walia, President and Chief Revenue Officer, Tata Communications, said: "AI has become one of the defining business priorities of our time, but the real differentiator is no longer AI itself-it's the infrastructure and integration that enable AI to deliver value at scale. Our research shows that while enterprise ambition is accelerating, readiness remains uneven. The organisations that will lead in the years ahead are those investing in the foundations that connect people, systems, data and intelligence across the enterprise.
"AI is a tightly coupled ecosystem of compute, power, connectivity and platforms, which are no longer independent systems-they are becoming one co-ordinated infrastructure. AI is accelerating this convergence, which Tata Communications addresses through its digital fabric of solutions. It is where we are uniquely positioned to enable customers to achieve their business goals."