Every team using AI already has a division of labour between people and AI. In most teams, nobody designed it.
Each person makes their own calls, day by day. The result is a boundary that nobody owns.
The third tool in The Augmented Us framework fixes that. It is a three-column document, reviewed every quarter.
The three columns
Column one: AI speeds it up. Work where AI makes the team faster or more consistent without weakening what the team is for. People supervise and edit. The team must be able to explain why AI here does no harm.
Column two: people lead. Work where the judgement, relationship or standard is the work itself. AI may supply research at the start. The decision, conversation and commitment stay human.
Column three: truly shared. Work where neither people nor AI could deliver alone. It needs a clear protocol: who does what, in what order, with what handovers and checks.
Why three columns
A simple split between AI work and human work fails. Most knowledge work is collaborative, and AI is now part of the collaboration.
The third column holds most of the value and most of the risk. It is where leadership attention pays back most.
Deloitte's 2026 Global Human Capital Trends makes a similar point. It calls for a shift from "human + machine" to "human x machine", with work redesigned for both to contribute.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
What goes wrong without it
Quiet substitution. Work that should be human drifts to AI. By the time anyone notices, the skill has faded.
Human in name only. Work labelled as human-led is mostly done by AI with light editing. The label becomes a risk.
Unmanaged shared work. Shared work happens without a protocol. Each person handles handovers differently. Errors slip through where neither the person nor the AI catches them.
Why it pays
Deloitte found organisations that prioritise deliberate work design are twice as likely to exceed their AI return expectations. Yet only 14% of leaders say they are good at shaping how people and AI work together.
This tool is a practical form of Adaptation: people directing, combining and checking AI in real work.
How to use it
Run a 90-minute session every quarter. List every type of work the team does. Place each in a column, with one sentence of reasoning.
Then compare with last quarter. Movement between columns is the most useful signal.
The leader holds the standard. Nothing enters column one without a clear reason. Nothing enters column three without a written protocol.
Try this
Place five types of work from this week into the three columns. Ask a colleague to do the same without seeing yours. The disagreements show where your division of labour is being decided quietly.
To check whether your team has time to review shared work properly, use the Checking Capacity Calculator.
Get the full framework. This article draws on The Augmented Us white paper. Download the white paper or take the team diagnostic.