Originally published in Channel Life

A significant portion of Australian organisations may be facing a problematic rollout of Generative AI technologies, according to the 2024 Data & AI Edge survey by technology research and advisory firm ADAPT.

The study, which surveyed 173 Chief Data & Analytics Officers (CDAOs) representing organisations responsible for over 35% of Australia’s GDP, highlighted a concerning lack of data maturity, resources, and skills necessary for the successful deployment of Generative AI.

Gabby Fredkin, Head of Analytics and Insight at ADAPT, expressed apprehension about companies moving forward with AI implementations without addressing fundamental requirements. “Companies need to seriously ask themselves how they hope to succeed with GenAI when the fundamentals haven’t been addressed. The vast majority of us still lack what’s needed to realise meaningful value from the technology: data literacy across our workforce remains extremely low, data infrastructure is immature, and data governance strategies aren’t anywhere near as robust as they should be,” said Fredkin.

One of the core issues identified by the survey is the lack of clear use-cases for Generative AI among almost half of the organisations surveyed. Despite this, 27% of companies intend to build, self-train, or host their own Large Language Models within the next 12 months. Fredkin pointed out that organisations with defined use-cases are primarily focusing on automating existing processes, yet he stressed the majority are unlikely to tap into Generative AI’s full potential.

“The tasks currently being performed by Generative AI have been the domain of machine learning for years, which suggests only a tiny few organisations are using the capacity of Generative AI to actually create things. If companies want the really exciting stuff to work as intended, they need buy-in from execs, who must be able to make informed, data-driven decisions, and end-users equipped with the know-how and guardrails to safely and effectively use the tech,” Fredkin observed. “Without it, I’m afraid many are on a collision-course with failure.”

The survey also revealed a strong correlation between AI readiness and data maturity. According to the study, the organisations best prepared to harness AI have, on average, nine times more highly-data literate employees. Additionally, a mature data architecture makes AI readiness seven times more likely, whereas robust data governance quadruples the chance of successful AI adoption.

Resource constraints pose another significant challenge, with 44% of data chiefs citing insufficient resources to execute their data strategies. Furthermore, two-thirds of companies face critical skill gaps in AI model engineering, and half are in dire need of data architects. The study suggests that overcoming these hurdles is crucial for realising the anticipated benefits of AI technologies.

Despite these challenges, Fredkin noted a silver lining, emphasizing the growing confidence among Australian data chiefs in their capacity to improve data strategies compared to the previous year. “There’s a silver lining to all of this. Australian data chiefs are much more confident in their ability to improve data strategies than they were twelve months ago and the issue of AI deployment has the attention of the C-suite,” he said.

Fredkin urged organisations to invest in modernising and simplifying their data and information architectures swiftly. “The message to them is this: You must invest in modernising and simplifying your data and information architecture as quickly as possible, because the firms who have already taken the modernisation issue seriously are out-earning their competitors, winning and keeping more customers, making use of better processes, and have happier employees working for them,” he stated. “I’m encouraged by the momentum that exists, but whether or not many in the economy are able to get up to speed in time remains to be seen.”