Deloitte finds more Australian firms linking AI tools directly to revenue

More Australian organisations are turning artificial intelligence into measurable sales rather than just internal time savings, according to new research published on Monday by Deloitte.

The consultancy firm surveyed 3,500 business leaders across 25 countries for its latest State of AI in the Enterprise report. In Australia, 27 per cent of respondents stated their AI deployments now generate revenue gains, a rise from 20 per cent recorded in January. Looking ahead, two-thirds of surveyed Australian organisations named revenue expansion as their top expected return from the technology.

Tools shift from advice to execution

The findings show a clear operational shift away from using machine learning purely as a drafting or research assistant. Today, 38 per cent of Australian companies let AI carry out decisions or operational actions without manual sign-off on a moderate level, although none reported full, unmonitored autonomy across the board.

Work patterns are also bending to fit these systems. Around 39 per cent of domestic firms are actively redesigning core workflows around software tools rather than bolting algorithms onto legacy routines. However, governance remains uneven. Almost one in five respondents admitted that internal accountability remains uncertain whenever an automated system makes a mistake or delivers a negative outcome.

Financial oversight is another area lagging behind rollout speeds. While three-quarters of local companies track their top-line software fees and application programming interface bills, only 38 per cent monitor actual token traffic, and merely one in four measures cost per individual query.

What this means for Australian businesses

Small and medium business operators can draw several practical lessons from how larger enterprises are structuring their investments.

• Look beyond basic time savings. The quickest wins often come from customer-facing workflows, such as lead scoring, quoting, and catalogue search, where faster responses convert directly into sales.
• Assign a clear human owner to every automated step. If an AI agent sends an incorrect order, mistimes a communication, or calculates the wrong price, a named staff member must hold responsibility for auditing and correcting the process.
• Watch usage metrics before monthly bills arrive. Because commercial AI services charge based on volume, tracking how many tokens or automated queries each tool consumes will prevent surprise cloud expenses as usage spreads across staff.