AI agent adoption rises among IT security leaders in ANZ

IT security leaders in Australia and New Zealand see considerable promise in adopting AI agents to address key security concerns, according to results from the Salesforce State of IT report.

The survey of over 2,000 enterprise IT security leaders globally, including 100 from Australia and New Zealand (ANZ), found that every respondent identified at least one security issue they believe could be improved with the use of AI agents.

According to the report, 36 per cent of IT security teams in ANZ have already integrated AI agents into their daily operations. This figure is expected to nearly double to 68 per cent within the next two years, signalling a significant upward trend in AI agent deployment across the region.

While optimism around the benefits of AI agents is evident, the report also highlights several challenges facing the region's IT security leaders. Notably, 58 per cent of ANZ respondents expressed concerns that their data foundations are not robust enough to fully realise the benefits of agentic AI. An equal proportion are not fully confident that appropriate safeguards and guardrails are in place for the safe deployment of these agents.

Despite these concerns, there is momentum towards greater adoption. Security leaders anticipate a range of benefits from AI agents, from enhancing threat detection capabilities to providing more robust auditing of AI model performance. Alice Steinglass, Executive Vice President and General Manager, Salesforce Platform, Integration, and Automation, said: "Trusted AI agents are built on trusted data. IT security teams that prioritise data governance will be able to augment their security capabilities with agents while protecting data and staying compliant."

The survey also indicates that up to 75 per cent of IT security leaders in Australia and New Zealand acknowledge that their current security practices require transformation. While 74 per cent see AI agents as offering opportunities to improve compliance, such as with privacy laws, nearly 83 per cent also identified ongoing compliance challenges as a key concern.

Cloud security threats, insider risks, and data poisoning—where malicious actors compromise AI training datasets—were listed among the most pressing risks for IT security leaders. In light of these evolving threats, 71 per cent of ANZ organisations expect to increase their security budgets in the coming year. This is only slightly below the global average of 75 per cent, reflecting a broad trend towards higher investment in IT security resources.

Complex and changing regulatory environments continue to pose challenges for IT teams aiming to deploy AI agents. Only 48 per cent of ANZ IT security leaders said they are fully confident that they can deploy AI agents in ways that are compliant with relevant regulations and industry standards. Furthermore, 85 per cent of organisations have not yet fully automated their compliance processes, leaving room for error and inefficiency.

There are signs of readiness and confidence among ANZ security leaders when it comes to their security and compliance practices. Sixty-one per cent believe their organisations are prepared for the development and implementation of AI agents. This figure is higher than both the Asia-Pacific average of 57 per cent and the global average of 47 per cent, indicating the region's relatively strong position in preparing for the AI era.

Maintaining effective data governance is seen as crucial to enabling the successful adoption of AI agents. Nevertheless, more than half of IT security leaders in ANZ are not confident that their organisations have adequate data quality or the right infrastructure required for AI deployment. Globally, CIOs report that budgets for data infrastructure and management are four times greater than those for AI itself, suggesting a focus on laying solid foundations before wider adoption of AI tools.

Building trust remains central to the deployment of AI within enterprises. Recent research indicates a decline in general consumer trust, with three-quarters of Australian consumers stating they trust companies less than a year ago. Additionally, 69 per cent believe advances in AI make trust increasingly important. Among ANZ security leaders, 59 per cent have not fully established ethical guidelines for AI use, 68 per cent lack complete confidence in the accuracy and explainability of AI outputs, and 58 per cent do not provide full transparency in the use of customer data for AI purposes.

Arizona State University is among the first higher education institutions to implement Salesforce's Agentforce digital labour platform, using trusted autonomous AI agents in operational workflows. The university has placed particular attention on data relevancy as it broadens its AI initiatives, and has adopted Salesforce-acquired backup, recovery, and archiving solutions to address its data management, compliance, and innovation needs.

The findings of the Salesforce State of IT report are based on a double-anonymous survey conducted between December 2024 and February 2025 across a broad range of countries including Australia, New Zealand, the United Kingdom, the United States, and others. One hundred respondents represented Australia and New Zealand.