According to a new finding from Pagerduty's AI Resilience Survey, Australian executives show the highest level of trust in the deployment of artificial intelligence (AI) agents within their organization.
The study was conducted through 1,500 IT and business executives in Australia, France, Germany, Japan, the UK and the US to examine how companies are leveraging AI-powered agents, the extent of executive trusts for such systems, and the emerging challenges associated with operational complexity and revival.
Highly trusted in AI agents
Research shows that 88% of Australian business and IT leaders trust agent AI to act autonomously on their behalf during a crisis, such as service outages and security events. This level of confidence is the highest in the world along with the UK. This study highlights significant changes in management attitudes, as AI is viewed as an essential infrastructure rather than as an experimental technique.
This trend is reflected in broader executive sentiment, with 81% of management investigating and saying they trust AI agents to take action on behalf of the company in crisis situations. In all regions, 74% of executives say their organizations struggle to function without AI.
“AI is at the heart of today's business operations, and future-looking leaders are implementing key use cases that could impact the organization's real business,” said David Williams, Senior Vice President of Products at PagerDuty. “Companies that embed automation and agents in their businesses will drive AI more efficiently, reduce costs, and strengthen customer trust. Pagerduty's latest research further shows that as recruitment and trust gains momentum and drives business outcomes, organizations are increasingly dependent on AI.”
Recruitment and integration in Australia
Australian companies have been shown to be pursuing the deployment of Agent AI at a rapid pace. The data shows that 88% of Australian organizations implement multiple AI agents, while 22% report launching such agents into business processes. Furthermore, 83% of Australian executives say they find companies that are difficult to operate without AI to be challenging.
Software development emerged as a key area of AI applications, with 88% of Australian business leaders using AI to write, review, or suggest code. Another 83% are actively testing AI-generated code. This is in line with international trends in which more than four out of five companies report similar use in coding practices.
Operational complexity and resilience
Adoptions are high, but this study highlights concerns about increasing complexity. Among companies with at least one AI agent, 84% of Australian respondents believe the complexity introduced by AI outweighs their ability to quickly and effectively manage. This concern is also reflected internationally, with 79% of organizations deploying multiple agents, and we expect that AI-driven complexity will soon exceed the number of people available to oversee and manage these systems.
Guardrails and risk management practices do not meet adoption. This report shows that we look at the global need for improved steps to detect errors or failures in AI tools worldwide. French executives are particularly concerned, with 90% acknowledging the need for better protection measures.
AI suspension and mitigation
Building resilience is a priority. 85% of Australian companies report experiencing at least one AI-related outage. Nevertheless, there is a high degree of confidence in its ability to detect and mitigate such obstacles, with 96% of Australian respondents expressing their guarantees on preparation. Globally, organizations are also preparing for potential issues. This is because more than half of those who have not yet experienced a halt have protocols in place to deal with such events.
Management perspective and future challenges
Executives attribute increased confidence in AI agents to a better understanding of improved output quality (49%), positive outcomes due to increased use (48%), how AI works (47%), and implementation of stronger monitoring measures (45%). Nevertheless, hands-on experience in deploying AI agents has made Australian companies more aware of the unique limitations of navigating the complexities associated with AI integration.
The study points out that there is a disparity in the consistency of testing AI-generated code. 85% of respondents report testing with AI-generated code, but only 39% do so consistently through the formal process. Regional differences were observed, with the US leading the formal test at 59%, while Japan advanced at 19%.
The findings suggest that as organizations advance adoption of agent AI, they are focusing on establishing better checks, procedures, and resilience plans to ensure operational continuity and robust monitoring of AI-driven processes.
