Pressure to implement AI plans is on the rise, but the readiness of enterprise networks to handle AI workloads has actually declined over the past year, according to new research from Cisco.
Cisco’s second annual AI Readiness Index notes a “huge chasm” between the urgency companies feel to deploy AI and their ability to actually do it. Some 8,000 global companies were surveyed for the AI Readiness Index, which aims to measure how prepared organizations are to invest in, deploy and use AI.
“Nearly all companies (98%) report the urgency to deploy AI has increased in the last year,” Cisco wrote. However, between 2023 and 2024, global AI readiness in the enterprise has declined. Today, only 13% of companies are fully ready to capture AI’s potential – down from 14% a year ago, Cisco stated.
Despite the focus on and investment in AI, business leaders do not feel they have made enough progress towards their AI ambitions.
“This is not deterring them, as leaders say they will not only continue to invest in AI, but actually increase their spend,” Cisco stated. The Index showed that 50% of those surveyed have between 10% and 30% of their current IT budget dedicated to AI.
“Interestingly, a large number of respondents in our survey noted that their AI investments have not yet delivered the gains they expected. Nearly 50% of respondents reported not seeing any gains or gains below expectations in areas such as assisting, augmenting, or automating a process or operation,” the report stated.
Among respondents who reported their AI implementations across top priorities have fallen short of expectations this year, 59% believe the impact from AI investments will surpass expectations after five years.
“Regardless of where they are on their AI journey, organizations need to be preparing existing data centers and cloud strategies for changing requirements, and have a plan for how to adopt AI, with agility and resilience, as strategies evolve,” said Jeetu Patel, chief product officer at Cisco.
The research found low readiness levels regarding the actual infrastructure and networking gear required to handle large AI workloads – which is worrying, Cisco noted, “especially as 93% of respondents predict that the workload of their organizations’ infrastructure will increase with the deployment of AI-powered technologies.”
Meanwhile, 54% of respondents acknowledge their infrastructure has limited or moderate scalability and flexibility to accommodate these increasing needs.
“Systems are struggling to keep pace with accelerating AI development as 79% of respondents say they require further data center graphics processing units (GPUs) to support future AI workloads, up from 76% last year,” the report stated. “Similarly, 78% of respondents lack confidence in the availability of computing resources for AI workloads, up from 76% in 2023.”
Cisco CEO on infrastructure modernization
In Cisco’s most recent financial call with Wall Street analysts, CEO Chuck Robbins said the vendor is seeing enterprise customers take on infrastructure modernization.
“Even when they’re unsure about what AI applications they may deploy six, nine, 12, months from now, they do know that they need modern infrastructure to be prepared to do so, and we’re seeing them… invest to get ready for it,” Robbins said. “This has been the same pattern we’ve seen for a few quarters now.”
“I think that the enterprise is actually looking at two different things that are going on there,” Robbins said. “Enterprises are updating their infrastructure to prepare for AI, and then they’re preparing for pervasive deployment of AI applications.”
Customers are in the process of deciding where to run applications in the cloud and what infrastructure will need to be implemented to support AI, Robbins said.
More from the AI Readiness Index
Some other interesting findings from Cisco’s AI Readiness Index include:
- Cybersecurity issues become critical. The proportion of respondents with limited understanding of threats specific to machine learning has increased from 65% in 2023 to 67% this year. Likewise, best practices in managing access control to AI systems and data sets are increasingly under strain, with nearly three-quarters (72%) falling short of a robust posture in this area, up from 68% last year.
- Companies feel less ready to manage data effectively for AI initiatives. Nearly a third (32%) of respondents report high readiness from a data perspective to adapt, deploy and fully leverage AI technologies. Most companies (80%) report inconsistencies or shortcomings in the pre-processing and cleaning of data for AI projects. This remains almost as high as a year ago (81%). Additionally, 64% report that they feel there is room for improvement in tracking the origins of data.
- A lack of skilled talent is a top challenge across infrastructure, data, and governance. Only 31% of organizations claim their talent is at a high state of readiness to fully leverage AI. In addition, 24% say their organizations are under resourced in terms of in-house talent necessary for successful AI deployments, and 24% say that there is not enough talent available in their sector with the right skillsets to address the growing demand for AI.
- Effective AI governance has become more difficult. When asked about the comprehensiveness of their organizations’ AI policies and protocols, 31% of the organizations said they are highly comprehensive. In addition, 51% of respondents identified “the lack of talent with expertise in AI governance, law and ethics in the market” as a challenge in improving their readiness from the governance perspective.
- There has been a noticeable reduction in cultural readiness to embrace AI. A lack of receptiveness to AI’s changes has contributed to the decline in cultural readiness: Boards have become less receptive to embracing the transformative power of AI, with 66% of them being highly or moderately receptive, down from 82% last year, while 30% of organizations report employees are limited in their willingness to adopt AI or are outright resistant.
