TL;DR – Productivity growth is flat, teams are already dabbling in AI, but only a handful of firms are scaling it for real value. Here is what the data and the early wins tell us about moving from pilot projects to enterprise-level impact.
Labour productivity fell 1 per cent over the year to March 2025 even as total hours worked kept rising.
Economists call the shortfall “capital shallowing”; I call it the tool gap: we have added people faster than we have upgraded their tech stack, so output per hour stalls. Putting in longer days will not change that, but smarter capital will.
AI adoption is stuck in first gear
Most knowledge workers have tried ChatGPT, Copilot or Canva Magic Write, yet only about 14 per cent of Australian organisations have taken AI beyond scattered pilots. The rest are burning time on proofs-of-concept that never reach the balance sheet.
Where it is already paying off
| Organisation | Sector | Focus | Tangible result |
|---|---|---|---|
| Commonwealth Bank | Finance | Scam detection & fraud alerts | Customer scam losses down 70 per cent, preventing ≈ A$650 m; call-centre waits down 40 per cent |
| Lawpath | Legal tech | Self-serve legal assistant | 25 per cent fewer service enquiries; quote turnaround dropped from 3 days to ½ day |
| Brisbane Catholic Education | K-12 | Copilot for teachers | Educators saved ~9 hours per week for lesson planning and student support |
| THE ICONIC | Retail | Vertex AI search & discovery | Null searches fell from 5 per cent to near zero, lifting revenue 2.6 per cent |
Common thread: each project tackled a painful workflow first, then wrapped governance around the model.
From experiments to enterprise value
- Fix the process, then add the AI. Automating broken steps just accelerates chaos. Streamline, then augment.
- Invest in people as well as platforms. Upskilling and clear guardrails remove the trust gap that kills roll-outs.
- Measure hours returned and risks reduced. Count time reclaimed, errors avoided and customer satisfaction gains, not model counts.
- Govern for trust. Use Australia’s AI Ethics Principles as the baseline so adoption scales safely.
Agenda ideas for FY 25-26 planning
- Create an agile AI investment pool. Shift a slice of CapEx from hardware refresh cycles into a fund that backs quick-hit productivity pilots across business units.
- Frame AI risk as strategic, not just technical. The bigger danger is being left with legacy workflows while competitors automate.
- Add a “Responsible AI Value” KPI. Track leading indicators like hours returned to staff, error rates cut and customer-experience lifts to keep focus on outcomes, not shiny tools.
Ready to close the tool gap?
Let’s talk about turning pilot learnings into a 90-day game plan that returns real hours and dollars to your business.
Written in partnership with the LuminateCX AI & Productivity practice. All figures current as of July 2025 and sourced from public case studies and official bulletins.