Chris Parker, Vice President, Asia Pacific & Japan, Riverbed, says gap between perceived capability and actual readiness shows that many organisations are not fully aware of the work required to make AI projects successful – and may be overconfident.
As AI adoption accelerates, Australian companies are investing an average of AU$22.9 million annually in AI initiatives, according to a survey conducted with business and IT decision makers. Despite this substantial commitment, many projects are not delivering the expected results.
According to Riverbed’s 2024 AI & Digital Experience Survey, only 18% of AI projects exceed expectations, while 59% meet them, and almost one in four (23%) underperform.
Readiness is the core challenge for AI projects. While 83% of Australian leaders and IT decision-makers feel they are ahead of their peers in AI adoption, only 35% report being fully prepared to implement AI effectively.
This gap between perceived capability and actual readiness shows that many organisations are not fully aware of the work required to make AI projects successful and may be overconfident.
AI depends heavily on high-quality, well-integrated data, yet 42% of respondents highlight data quality as a barrier. Companies often underestimate how fragmented or incomplete their data sets are, which leads to flawed AI outcomes. In our survey, less than half of the organisations rate their data as excellent in key areas such as completeness, accuracy, and accessibility.
The issue isn’t just the availability of data but also the need for integration and visibility. For AI to deliver meaningful results, organisations must ensure their data is reliable, comprehensive and visible across the enterprise.
The human factor is equally important. Despite concerns that AI may reduce the number of jobs, the reality is that it requires skilled professionals to manage, deploy, and scale it. However, these skills are new and not widely available.
Many professionals in the market still lack the expertise needed for AI success, and fewer than half (49%) of large Australian companies provide extensive training on AI. The lack of skilled talent can significantly contribute to underperforming projects, as even the most sophisticated AI tools can fail to deliver without the right expertise.
How businesses can improve outcomes
Companies need to focus on several key areas to improve the success rate their AI projects.
They must build strong data foundations capable of working on an AI context. Organisations should prioritise data quality and integration, using tools that improve data observability to monitor and resolve issues early. This ensures AI systems work with the most accurate and complete data possible.
Additionally, prioritising real-world data over synthetic data is important, as it better captures the complexities necessary for effective AI training.
Companies also need to invest in their people. AI tools are not plug-and-play solutions; they require knowledgeable teams to manage and refine systems. Organisations should invest in continuous AI training and education to close the skills gap, ensuring that their workforce can effectively leverage AI.
Planning and project clarity are other prerequisites for achieve good results. Many AI projects falter because organisations don’t set clear, realistic goals or fully understand what success looks like. Companies need to ensure that AI initiatives are grounded in well-defined business outcomes and are supported by a clear roadmap that includes scalability, governance, and ethical considerations.
With 79% of companies worried about AI accessing proprietary data, businesses also need governance frameworks to mitigate security risks and ensure compliance, implementing privacy and security protocols. These frameworks help ensure that AI projects are both secure and scalable, enabling long-term success.
Our data shows that investments and optimism are critical but not enough. Success depends on how companies address these underlying challenges. AI’s potential is real, but to unlock it, organisations must approach the technology with a clear, realistic understanding of what it takes to make it work.
