AI ROI in 2025: Stop Counting Hours, Start Finding Focus

The wrong question

In many executive leadership meetings across Australia and APAC, the question most leaders are still asking is: “How many hours will AI save us?”

It’s the wrong question.

The reality in 2025 is that AI isn’t automatically giving people time back. In fact, many employees using AI tools are reporting longer workdays, less focus time, and more stress (ActivTrak, 2025). This is the AI Productivity Disconnect, where the very tools designed to make us more efficient are creating new layers of noise, context switching, and rework.

Hours saved is a vanity metric. Focus returned is the measure that matters.

Incremental, not instant

AI ROI in 2025 is showing up in small, compounding gains, not dramatic overnight transformations.

Case studies across APAC highlight this.

Healthcare: AI scribes reduce per-patient documentation time by up to 15%, but unless workflows are redesigned, after-hours admin can increase (RACGP, 2025).
Education: AI marking tools cut first-pass grading time by nearly 50%, but final moderation still requires humans (Marking.AI, 2025).
Enterprise pilots: Median reported ROI sits at ~10%, well below target rates, because “time saved” often isn’t reinvested in higher-value work (BCG, 2025).

This mirrors the CRM and ERP era. Back then, value didn’t come from the software install but came years later, after process maturity, cultural change, and organisation-wide training (Masood, 2025). AI is following the same curve.

What works in APAC

The organisations breaking through the disconnect are the ones treating AI as an organisation-wide program, not a point tool.

Telstra: Their governance model includes an AI model register, cross-functional risk councils, and direct board oversight. This structure doesn’t slow them down—it removes ambiguity, so projects move faster with clear guardrails (UTS, 2024).
National Australia Bank (NAB): Their “Data & AI Month” has reached 10,000+ employees with structured learning, reducing fear and normalising AI as a safe, sanctioned capability (NAB News, 2025).

Both prove the same point: AI ROI flows when you invest in governance and skills as much as the tools themselves.

New KPIs for a new era

If “hours saved” is dead, what replaces it?

The most forward-thinking enterprises are adopting human-centred KPIs alongside financial metrics (Worklytics, 2025):

Focus-Time Fragmentation Index – interruptions per block of planned deep work (target <1.0).
Time Reallocation Ratio – % of “saved” time demonstrably reinvested in higher-value work.
Decision Speed – cycle time from problem to recommendation.
Adoption Quality – % of users on sanctioned tools, competency-verified.
Healthy Work Patterns – % of staff maintaining sustainable workloads.

Boards don’t need another 30-page AI report. They need a one-page AI ROI dashboard that blends financial, operational, human, and customer metrics into a credible story they can trust.

The opportunity for leaders

Here’s the truth: AI ROI is not about chasing instant transformation. It’s about designing for incremental, organisation-wide improvement, measured honestly and reinvested wisely.

If leaders keep chasing “hours saved,” they’ll miss the dividend. If they track focus, flow, and outcomes, they’ll unlock it.

Call to action

If you want to see where AI can deliver meaningful productivity in your business, book a session with me.

We’ll run the Effort vs Impact Productivity Report with your teams to pinpoint bottlenecks, reduce fragmentation, and prioritise the next two or three moves that deliver incremental productivity gains that stick.