McKinsey’s 2025 AI Report: Wide Adoption, Limited Scale. What Australian Leaders Must Do Next
McKinsey's 2025 survey shows wide adoption but limited scale. Leads must choose ambition or be left behind.
"AI isn't an IT project, it's a leadership test" - Stephen Hunt, The Square Wave
McKinsey’s State of AI in 2025 Report confirms what many Australian leaders feel in their bones: AI is everywhere in the business, but real enterprise value is still rare. 88% of organisations report regular AI use, yet only about one third have moved beyond pilots to scale. Agentic systems are creating fresh opportunity and fresh complexity, with roughly six in ten organisations experimenting and nearly a quarter beginning to scale agents in at least one function. The ‘so what?’ for leaders is straightforward: this is the moment in time to shift from tinkering to purposeful transformation.
Read on for the executive implications you can act on this quarter to protect revenue, accelerate growth and shape workforce change.
If you prefer to listen, check out this AI-generated podcast made with NotebookLM discussing the content:
For a quick explainer, check out this AI-generated video summary made with NotebookLM:
Broad adoption but shallow embedding
The report shows broad adoption but shallow embedding. Use is spreading across functions (IT, marketing, knowledge management and service operations show especially high activity), yet enterprise-level EBIT impact remains limited: only 39% of respondents attribute any measurable EBIT effect to AI, and most of that is under 5%. Larger firms are more likely to scale, and a small group of “high performers” (about 6% of respondents) capture disproportionate value by aiming for growth and transformation rather than cost only. Agentic AI is no exception: many organisations are experimenting, but few have integrated agents into multiple functions at scale.
Why this matters for Australian leaders
For Australian senior leaders that matters for three reasons:
Competitors who pair ambition with disciplined operating practices will widen the performance gap.
Half-measures create risk: partial automation without workflow redesign increases error, reputational and compliance exposure.
Workforce impacts are uneven: while a plurality expect little near-term headcount change, a meaningful minority anticipate cuts or role shifts. Leaders must therefore balance urgency with governance and a value roadmap that is shared with employees.
Outcomes follow choices
The clear pattern in 2025 is that outcomes follow choices. Organisations that see meaningful EBIT impact share a set of behaviours: bold ambitions for AI-driven innovation, senior leader ownership, deliberate workflow redesign and investments in people, data and platform. High performers are more likely to set growth or innovation objectives alongside efficiency goals, to commit larger shares of digital budgets to AI, and to use human-in-the-loop practices to manage risk. They are also three times more likely to have senior leaders who actively champion AI adoption.
From side project to strategic capability
For Australian boards and executives the takeaway is simple: AI will not be a side project. To capture value you must treat AI as a change management program and a capability program, not an IT project. That means defining the outcomes you want from AI, aligning investment and incentives to those outcomes, training your people to enhance their abilities with AI and embedding the right operating model to scale. It also means being pragmatic about agents: use them where they materially shorten cycle time or unlock new customer value, and test them under clear guardrails before expanding.
Three to four actions your executive team should prioritise now
Set dual objectives. Define at least one growth or innovation target for AI while considering efficiency targets as a secondary objective so they don’t impact culture negatively. High performers will set both and see broader enterprise benefits as they skilfully lead their people through the change.
Fund a sprint to redesign one core workflow. Pick a high-value process, map current handoffs, and rebuild it with AI and human roles clearly specified. Evidence shows workflow redesign is a top predictor of impact.
Appoint an accountable leader and aim for rapid governance. Senior ownership matters. Assign a visible executive sponsor, create decision rules for agent deployment and define human validation thresholds for model outputs. This reduces risk and speeds scale.
Invest in measurable capability. Allocate training, hire a small set of critical roles (data engineers, platform leads), and track KPIs that link AI work to revenue, cost and customer outcomes. Measure adoption, not just experiments.
These steps move you from experiments to sustainable value capture. They are not purely technical tasks; they require strategy, change leadership and clear metrics.
Australia’s AI moment
Australia’s AI moment is not about chasing every new tool. It is about choosing where AI becomes a business differentiator and then resourcing that choice properly.
The Square Wave is on hand if you need support to guide your 'human-centred AI training and transformation' project. We can help train your teams, reimagine your workflows, craft strategic roadmaps, draft functional playbooks and explore governance frameworks tailored to Australian regulation and markets.
If you want to compare notes with peers, join the AIIAA network to share best-practice and influence standards for ethical AI implementation. Together we can ensure AI strengthens workplaces and creates competitive advantage across industries.
Let's go! 🚀
Steve.
