Australia’s major financial institutions are accelerating their transition from experimental generative AI into enterprise-scale agentic systems, with cloud infrastructure emerging as the critical governance and execution layer that will separate winners from laggards in the race for measurable returns.
An interesting set of data from Capgemini Research Institute’s World Cloud Report for Financial Services 2027, released this month found that 59% of financial institutions globally are already piloting or deploying multi-agent AI systems in production environments.
Yet the ambition is running well ahead of execution, with only 10% of firms having achieved the enterprise-wide orchestration necessary to run autonomous multi-agent workflows across multiple business lines simultaneously.
For Australian financial services firms, this orchestration gap is particularly consequential. The country’s four major banks, along with Macquarie and the leading insurers, are navigating a simultaneous set of pressures; namely the need to demonstrate commercial returns on AI spending, the obligation to satisfy an increasingly assertive regulatory environment, and the technical challenge of modernising legacy infrastructure that in some cases dates back decades.
AI Investment Shift Underway
On the capital allocation side, the shift is already underway, and the study data projects that executive prioritisation of outcome-based ROI funding will rise 141% by 2028, while simpler efficiency-focused spending is expected to fall 42% over the same period.
Top Australian Brokers
- Pepperstone - Top Australian broker - Read our review
- eToro - Invest in ASX and international shares - Read our review
The implication is that the era of exploratory AI projects funded on the basis of strategic positioning alone is drawing to a close. Boards and CFOs are demanding demonstrable bottom-line impact, and roughly half of all financial institutions currently lack the measurement frameworks to provide it convincingly.
Model Routing
Cost discipline is also reshaping the technology stack, with ~90% of financial firms planning to implement intelligent model routing architectures. Token costs and compute expenses have emerged as a meaningful constraint on the scalability of agentic systems, making model routing all the more important.
Coming back to the report, 52% of banks and insurers are now prioritising major hyperscalers, specifically AWS, Microsoft Azure, and Google Cloud, as their preferred execution authority for agentic AI.
The selection criteria are dominated not by raw capability but by regulatory compliance, data residency requirements, and the availability of certifications that satisfy APRA and ASIC expectations. Cloud is no longer simply a cost-reduction tool but the data spine, control plane, and governance fabric through which autonomous agents must operate.
Sourcing Strategy
In response to these competing pressures, the sector is converging on a hybrid sourcing model. Rather than attempting to build all agentic capability in-house or relying entirely on packaged vendor solutions, more than two-thirds of financial institutions are expected to adopt a three-track approach within three years.
That would involve building proprietary systems for core IP and customer-facing differentiation, buying pre-built agentic software for routine operational workflows such as tax processing, compliance checking, and reconciliation, and partnering with ecosystem providers for platform depth and cloud governance infrastructure.
The research identifies a cohort of global frontrunners, representing roughly the top 5% of financial institutions by AI maturity, that have already embedded this structure. These firms consistently outperform peers on cross-functional orchestration metrics and demonstrate more reliable compliance with cloud governance requirements.
Financial institutions are now at a point where the strategic promise of agentic AI is colliding with the practical realities of legacy infrastructure, governance gaps, and a regulatory environment that is tightening on a defined schedule. The Tier-1 banks with the deepest compliance capabilities and the most advanced cloud partnerships are best positioned to convert that promise into durable earnings growth, but there are plenty of hurdles to overcome.
What this all means for ASX financial stocks will be told in time, but with AI investment continuing at pace, firms will need to find ways to avoid silos across the business in order to truly extract maximum value.
