The Australian Market Scenario for AI in 2024​

Overview:

The AI hype cycle remains robust, with many tech vendors, both large and small, advertising their products as “AI-enabled.” It is understandable that Australia’s C-suite executives might question whether the enthusiasm surrounding these tools translates into tangible business benefits

Enterprises are taking a conservative approach to deploying AI, choosing to “test and learn” when rolling out, for example, generative AI tools based on large language models (LLMs).

Australia’s most valuable company, the Commonwealth Bank, is trialling ChatGPT-style software with internal staff before it goes live with the tool across its large contact centre.

At the other end of the spectrum, humanitarian organisation, the Australian Red Cross, has launched a gen AI bot to assist workers with basic tasks during support calls. However, gen AI is not being used by volunteers and staff in the field because of its potential to make decisions that could negatively impact people in need.

The efficacy of these LLMs for mission critical applications remain uncertain due to their current level of low maturity.

Key Takeaways:

• In February, ADAPT surveyed 290 Australian CIOs about their readiness to harness AI in the next 12 months, with only 9% indicating they were prepared. By August, this figure had risen to 30%, demonstrating significant progress among enterprises. CIOs who felt they were unprepared fell dramatically from 66% to 22% over the period.
• AI and data initiatives across organisations seem to have deprioritised in favour of efforts to modernise core tech infrastructure aimed at reducing costs.
• Regulatory changes that will increase pressure on organisations to uplift their data foundations to support responsible AI activities are yet to be defined.
• The top five AI use cases for Australian CIOs are productivity or efficiency, innovation or disruption, opportunity, hype and augmentation. For chief data and analytics officers (CDAOs), the five use cases are process assistants, productivity or efficiency, insights enablement, customer enablement, and document search.
• The top barriers to AI success in August are data quality issues, stringent or unclear regulatory requirements (unsurprising), insufficient data skills, adverse culture and user resistance, and a lack of clearly defined use cases.
• Large numbers of organisations are immature when it comes to building mature data foundations that will support responsible AI deployments. CIOs surveyed say their organisations are immature across multiple areas including capturing and storing data, building a robust information architecture, removing data silos, as well as creating a single view of the customer.
• When deciding on how to resource their AI activities, most organisations will either source and retain capabilities in-house or use a mix of internal and outsourced tools and talent.
• $4.7 million or 11% of the CDAO’s average total annual budget of $43 million is spent on artificial intelligence tools and capabilities.
• AI risks caused by external software developers are considered severe by 40% of 1o3 Australian CISOs surveyed by ADAPT and moderate by an additional 52%.
• Insufficient human oversight of computer-generated code is viewed as a severe risk by 37% and moderate by 53%.
• The use of open-source software libraries is deemed a severe risk by 37% and moderate by 52% of respondents.

Download these slides to learn more.