Australian organizations report widespread difficulty proving the business value of AI investments, according to an APAC survey by Ecosystm, an analyst firm focused on data preparation and AI implementation.
The survey was commissioned by Snowflake and was based on responses from over 700 business and IT leaders in the Asia Pacific and Japan regions. We investigated how organizations apply agentic and generative AI, where they evaluate use cases, and what impedes progress from pilot projects to widespread deployment.
In Australia, 81% of respondents said it was difficult to demonstrate business value and return on investment from AI. This number ranks second among the surveyed markets, behind South Korea’s 83%. In New Zealand, 70% reported having similar challenges, the lowest of any country surveyed.
Customer use case
The study found that Australian organizations most frequently evaluate AI for customer-facing interactions. Approximately 67% said they have evaluated use cases for interacting with customers across channels. A further 53% cited improved chatbot responses. Another 65% cited marketing content creation.
New Zealand respondents also prioritized customer interactions across channels, with 68% responding. Approximately 60% reported evaluating AI for marketing content generation. An additional 56% cited improved search and summarization of data and reports.
This study describes the gap between early experimentation and consistent delivery of business results at scale. The gap had more to do with the data and technology infrastructure than with the AI model.
“Enthusiasm for AI is growing across the region, particularly in Australia and New Zealand, and business leaders now expect this technology to deliver real business value,” said Theo Urumuzis, Snowflake’s senior vice president for Australia, New Zealand and ASEAN.
“To unlock this value, organizations must deeply integrate AI into their business strategy, rather than pursuing it as a mere experiment. To do this, they must start with clear, measurable use cases tied to real business needs, rather than deploying AI for AI’s sake.”
data failure
Across Australia and New Zealand, respondents cite data accessibility as their biggest data challenge. Approximately 56% of respondents reported this issue. Data quality followed at 52%. Data security and data observability were both 49%.
The study also found that the level of integration of AI into overall business strategy is low. Only 19% of Australian businesses say they have fully integrated AI into their business strategy. In New Zealand, 24% reported full integration.
The report cites data fragmentation and poorly prepared technology infrastructure as common constraints to AI adoption. He also pointed out the difficulty of handling unstructured data at scale across an organization.
According to the survey, 38% of organizations across countries surveyed said they had invested in technology to analyze unstructured data.
partners and platforms
The study also notes a focus on external support for AI programs. It found that 85% of Australian organizations have engaged or plan to engage with technology partners for strategic, technology and data needs related to AI projects. In New Zealand, 76% reported taking the same approach.
Hourmouzis has combined the results of AI with extensive implementation and governance efforts across the organization and its suppliers.
“AI is more than just plug-and-play. Partners are essential to closing capability gaps, accelerating deployment, and establishing governance frameworks to help organizations stay ahead of the next wave of disruption,” Holmzis said.
The study also identified a set of practices for measuring value over time. The report argued that while pilot projects can demonstrate feasibility, organizations may miss out on costs and benefits entirely if they implement AI into their operations. We noted the need to track infrastructure and model maintenance, governance and compliance considerations.
The study also addresses the impact of fragmented tools across the AI lifecycle. We argued that disconnected systems across data preparation, model development, deployment, and monitoring can obscure performance and cost. It discusses integrated measurement of technical and business metrics as a way organizations can approach visibility and accountability for their AI programs.
The report found that organizations in Australia and New Zealand are showing signs of moving towards a more strategic approach to AI programs, with partner engagement and investment decisions related to data infrastructure remaining a key focus in the next stage of deployment.
