AI on the Edges Cannot Fix a Foundation Problem
When AI arrived, the expectation was that it would dissolve this kind of structural busywork and redirect human effort toward work that actually requires human judgment.
For many ANZ organisations, that shift is still incomplete.
AI is delivering value, but too often from the edges of work rather than from inside the systems and workflows where work actually happens.
Earlier this year, we shared research that identified what we called 'the productivity tax': roughly 40% of the time AI was supposed to save is spent reviewing, fixing and reworking AI outputs.
Our new research shows exactly why.
Most organisations have bolted AI onto the edges – a writing assistant plugged into email here, a chatbot stuck in front of a legacy portal there.
The outputs emerge, and a human is still required to collect them, verify them, and carry them across to the next system by hand.
The workflow has not changed. The AI has simply generated more content that needs to be manually moved around within it.
Only 30% of ANZ employees say their organisation has embedded AI directly into their core workflows.
The majority are running AI around the work rather than inside it – and the consequences show up precisely where the data is already most troubling.
Seventy-three percent of ANZ employees say decisions are delayed when information is missing or unclear. Sixty-four percent say their teams regularly disagree over whose numbers are right.
Applying AI to the edges of this environment does not resolve those disagreements. It risks accelerating them.
As another IT Director observed during the research: "A significant portion of the day is lost in meetings where the adoption of AI is discussed, but the discussion consistently revolves [around] simplistic justifications rather than practical choices."
That is what happens when organisations skip the harder question. Not 'where can we bolt on AI?' but 'where does work actually fall apart between our systems, and can AI own that gap?'"