Governed choice: the key to scaling AI safely in Australian business

Australian businesses do not have an AI ambition problem. They have an execution problem. Increasingly, that challenge comes down to choice: giving teams access to the AI tools that fit their work, without adding more risk, cost, and complexity.

For many organisations, especially in the mid-market, the AI tools are spreading faster than the rules around them. The result is familiar: promising pilots, scattered tools, and limited business impact.

Recent research from Deloitte and PwC points to the same challenge. Many Australian organisations are struggling to turn AI into scalable outcomes. Deloitte has reported that only 12 per cent of Australian organisations say AI is transforming their business, compared with 25 per cent globally, while only 65 per cent plan to increase AI investment, compared with 84 per cent globally. PwC has similarly found that only around one in five Australian organisations report they have the data, governance, and technology foundations needed to scale AI confidently and safely.

For CFOs, that gap matters. AI's success in business ultimately comes down to the control, productivity, automation, and value it delivers. If AI is going to support forecasting, reporting, supply chain decisions, or customer operations, it must be reliable, auditable, grounded in business context, and connected to workflows.

Fragmentation and missing context

Many businesses have added AI tools across finance, operations, supply chain, and customer service without a unified strategy. That creates integration bottlenecks, data duplication, and inconsistent outputs, making it harder for leaders to see where value is being created, where risk is building, and whether outputs can be trusted.

Employee enthusiasm for generative AI has also fed the rise of shadow AI, where employees use unapproved consumer-grade tools to speed up work. While these consumer tools can be useful, they can also expose sensitive data, create intellectual property risk, and generate outputs that are difficult to validate.

In addition, generic AI tools sit outside core business systems, removed from the business context of the request. To enhance accuracy, insights, and decision-making, AI needs to have access to contextual data such as financial structures, approval hierarchies, operational workflows, and business rules. Without that context, even a polished answer can be wrong in ways that are hard to spot.

Embedded AI and MCP connectivity

The next phase of AI adoption will require flexibility and giving businesses different ways to adopt the technology.

AI should be embedded in core systems to improve productivity, reduce manual work, and help teams act on insight in context. At the same time, organisations should have the flexibility to bring their own AI models and tools into the business in a governed way when those tools are better suited to a specific use case.

True Protein, one of Australia's leading health and nutrition brands, offers a useful example of how AI can be adopted in different ways to achieve desired outcomes. With finance, inventory, manufacturing, warehousing, and sales connected in NetSuite, the company has reduced month-end close by 70 per cent, from 10 days to three, while gaining real-time visibility across production and fulfilment that helps teams move faster and reduce errors. True Protein says NetSuite's AI-powered features are also helping improve productivity and day-to-day decision-making, showing how embedded AI becomes more valuable when it sits inside the systems and workflows the business already relies on.

At the same time, the company is experimenting with popular large language models where it makes sense. For example, True Protein uses NetSuite AI Connector Service with NotebookLM and Gemini to turn general ledger data in NetSuite into podcast-style reports, helping leaders stay across the numbers on the move.

True Protein is a useful example of what governed choice can look like in practice: embedded AI working inside core workflows, alongside third-party models where they add value, all grounded in the same business data and controls.

Australia's mid-market organisations are often digital-first and ambitious. But they also need to balance innovation against risk and speed against control.

The companies that benefit most will be the businesses that use AI to make work more efficient, decisions more timely, and execution more consistent.