Australian lenders adopt AI faster than data readiness

Australian financial institutions are adopting agentic AI in underwriting faster than they are preparing their data for it, according to Experian. Its research found 72% of surveyed institutions already use the technology to support underwriters.

Only 3% of respondents said their data was fully ready for AI-driven decisioning, while 67% said it was either not ready or only partially ready. The figures point to a widening gap between AI adoption and the quality, integration and governance of the data used in credit and fraud decisions.

The findings come from a global study of more than 800 senior decision-makers and over 80 expert interviews across 12 countries. In Australia, the study covered 102 respondents and focused on the use of AI, data and software in credit and fraud risk underwriting.

Many lenders appear to be using AI in limited or developing ways rather than across their full decisioning processes. More than seven in 10 Australian respondents, or 72%, described their organisations as emerging or early in their use of AI across fraud and credit risk underwriting. Only 11% said AI was widely implemented across underwriting processes.

Speed remains a central reason institutions are turning to these systems. Three in four respondents, or 75%, said faster or real-time decision cycles were among the core benefits AI, data and software could bring to underwriting operations.

Operational barriers

The study suggests the main obstacles are not interest in AI itself, but the practical challenges of using it reliably in production systems. Fragmented data systems that do not provide a unified customer view were identified as a key barrier by 45% of respondents.

Poor data quality followed closely, cited by 42% of Australian respondents. Another 31% said a lack of trust in AI outputs was a significant hurdle, underscoring the difficulty of moving from experimental tools to decision systems that can be relied on in credit and fraud assessments.

Governance also featured strongly. Some 69% agreed that data quality and governance were among the reasons AI implementations fail, while 84% said transparency of analytics and insights was highly valuable in improving decisions.

That emphasis on transparency reflects broader pressure on lenders as they balance faster decision-making with scrutiny over how customer data is collected, linked and used. In credit and fraud controls, explainability is becoming more important as institutions seek to show how an outcome was reached and whether human oversight remains in place.

Human oversight

The research indicates that many Australian financial institutions remain cautious about handing full authority to AI. More than half of respondents, or 51%, said they were comfortable allowing AI to make decisions without human review only in low-risk cases.

Just 2% said they were comfortable with fully autonomous decisioning at scale across most use cases. That suggests lenders are more willing to use AI for support, recommendations and triage than for unrestricted final decisions in high-stakes lending or fraud scenarios.

The level of caution is notable given the strong interest in broader adoption. More than nine in 10 respondents, or 92%, said they would pilot, test or adopt a vendor that could meet their data, software and AI requirements for fraud and credit risk underwriting.

The figures suggest the market for integrated decisioning systems will depend less on persuading lenders to use AI and more on helping them resolve persistent issues in data management, system integration and oversight. The challenge for suppliers and institutions alike will be to connect those elements in a way that supports faster decisions without weakening control.

Mathew Demetriou, Managing Director, Software Solutions, Experian Australia & New Zealand, said: "What we're seeing with Australian lenders is that AI is already in underwriting workflows, with the research showing 72% are using agentic AI for decision support. But the harder question is how to trust and govern it at scale, especially as regulators sharpen their focus on how data is used. The foundations underneath AI still need to catch up."

He added: "Data quality, integration, trust and governance may determine whether AI can move from contained use cases into core decisioning. Without trusted and well-governed data, there is a risk organisations may struggle to operationalise AI at scale. With appetite for integrated AI-enabled decisioning strong, the industry's next phase may be defined by how quickly it can close the data readiness gap by connecting data, AI and governance into a single, trusted decisioning environment. It's what we call connected intelligence."