AI readiness stalls in APAC

Despite widespread investment in artificial intelligence (AI) in Asia-Pacific, the region’s adoption of the technology is stalling, according to Cisco’s 2024 AI readiness index recently unveiled at Cisco Live Melbourne.

While 86% of respondents believe they have just 18 months to deploy AI before facing negative business impacts, only 15% feel truly ready today. This stark discrepancy underscores the mounting pressure to get AI implementations right, a challenge compounded by the fact that AI readiness has declined since 2023.

“The more you know, the more you realise you don’t know,” said Raymond Janse van Rensburg, vice-president of specialists and solutions engineering at Cisco Asia-Pacific, Japan and Greater China.

Half of the respondents in the study said between 10% and 30% of their IT budget was being spent on AI, but 40% said gains were non-existent or below expectations. Despite that, AI spending is expected to double over the next four or five years, said Carl Solder, Cisco’s chief technology officer (CTO) for Australia and New Zealand (ANZ).

The report highlighted several key hurdles in AI adoption. While 95% of respondents have an AI strategy in place, infrastructure readiness is lagging. A significant 58% want to prioritise the scalability and flexibility of their AI infrastructure. And while 79% said they need additional graphics processing units (GPUs), only 31% nominated that as a priority.

Data quality remains a major obstacle, with less than a third of respondents feeling their data is truly AI ready, underscoring the need for pre-processing and cleansing. Governance is another area of concern, especially given evolving data privacy laws and regulations across the region.

And the talent shortage, already acute across IT, extends to AI, with 45% citing it as a challenge. The lack of skilled talent, combined with rising hiring costs, is prompting organisations to invest in upskilling and reskilling (though just 43% are doing so), a necessity for wider AI adoption, according to Cisco ANZ vice-president Ben Dawson.

The cultural readiness component also reveals a worrying shift, with a significant drop in the number of organisations demonstrating a clear commitment to AI at all levels. This includes reduced support from the top, with less commitment from board level down to leadership teams. Furthermore, a concerning 29% of respondents reported resistance from employees at the grassroots level.

Despite the challenges, Cisco remains optimistic about the future of AI in APAC. In ANZ, it expects to see strong demand for AI-ready datacentres, fuelled by the decarbonisation drive in the energy sector and modernisation efforts in healthcare. Cisco is actively helping its customers navigate this complex landscape through a combination of product enhancements and strategic partnerships.

Cisco’s acquisition of Splunk has resulted in a powerful combination of observability tools, allowing organisations to monitor and optimise their digital experience across hybrid and multicloud environments. Earlier this year, the company also announced its Hypershield architecture, a distributed security solution designed to provide fine-grained protection across the entire network infrastructure, ushering in a new era of automation and minimising network outages.

Cisco is also making significant investments in AI-powered collaboration tools, like the new Webex Workspace Designer, which simplifies the selection of appropriate collaboration devices for various meeting room types and configurations. The company is also expanding its AI-driven security solutions, with a focus on improving the efficiency and effectiveness of security operations centres.

“AI does matter, regardless of the hyperbole, but you got to put guardrails around it,” said Tom Gillis, Cisco’s security business group senior vice-president and general manager, adding that it can be hard to achieve. For example, an application and a database are separate entities in a traditional architecture, but with AI, “the data is infused into the model – it never forgets”.

Furthermore, the separation of duties is an important part of security which needs to be replicated in an AI environment – if a chatbot has access to employee data, employees should not be able to persuade it to give them or their colleagues a pay rise, Gillis said.

Cisco’s continued commitment to helping customers achieve their AI goals is evident in its extensive work with key customers across the region. Energy Queensland, for instance, is using Cisco ThousandEyes to gain “hop by hop insights” into network issues, enabling proactive problem resolution before users even notice a problem.

The University of Melbourne, a research institution with a large and complex infrastructure, relies on Cisco’s technology to navigate the evolving security landscape and keep pace with the demands of cutting-edge research. The university has its own primary datacentre, and its fixed and wireless networks (with 6,000 access points, including some of Cisco’s new Wi-Fi 7 models) support 40,000 unique connections. “In effect, we’re running a small city”, and that does push the limits of the technology, said its CTO Byron Collins.

Cisco’s efforts are also focused on developing responsible AI practices, recognising the need for guardrails and ethical considerations. The company is actively involved in initiatives like the Digital Well-being Hub, a joint project with the OECD (Organisation for Economic Co-operation and Development), to study the impact of digital technologies on people’s lives. The initiative aims to collect data on online activity to understand the potential unintended consequences of excessive digital device or media use, particularly in the context of AI-powered applications.

“So many see AI as a threat” to their security or wellbeing, said Guy Diedrich, Cisco senior vice-president and global innovation officer, “but there’s no thought leadership in this space” because nobody is an expert yet. Cisco is involved, in part, because “we see things others can’t” as much of the world’s data traffic passes through Cisco equipment, he added.