
For your team
Team Practice
Module · Inside the work, or beside it
The Team AI Workflow Check
The gap between a team that got value from AI and one that just bought licences is almost always the same: whether the AI sits inside the daily workflow or off to the side in a separate tab. AI beside the work adds steps, creates handoffs, and quietly duplicates effort. AI inside the work removes friction and gives time back. This module checks the five things that decide which one you have: whether AI lives in the tools your team already uses, how much friction sits at the handoffs, whether AI and people are redoing each other's work, who owns the workflow, and whether you can actually measure the time it saves.
What the five levels look like
Every dimension in this assessment is scored 1 to 5. This is what the levels mean, dimension by dimension. The graded report diagnoses where your own answers land and what to do about it.
AI is in the tools
- 1Separate tab only
- 2Copy-paste between
- 3Some integration
- 4In the main tools
- 5Native to the workflow
At the low end: AI in a separate tab is AI your team has to remember to visit, which means most of the time they will not. Get it into the one or two tools where the work actually happens, and adoption follows. What good looks like: AI native to the workflow is AI people use without thinking about it, which is the whole point. Keep it embedded as tools change; an integration that breaks silently sends everyone back to the tab.
Handoffs are smooth
- 1Constant reformatting
- 2Rough handoffs
- 3Some friction
- 4Mostly smooth
- 5Seamless handoffs
At the low end: If every handoff means reformatting and reshaping, the AI is creating work at the seams even as it saves it in the middle. Fix the roughest handoff first; that is where the time is leaking. What good looks like: Seamless handoffs make AI feel like part of the team rather than a step to manage. Keep watching the seams as workflows change; new friction loves to hide at a handoff.
No duplicated work
- 1Heavy duplication
- 2Often redone
- 3Some overlap
- 4Little duplication
- 5Clean division of labour
At the low end: When people routinely redo what the AI produced, you are paying twice and saving nothing. Watch one workflow end to end and find where the effort is being duplicated, then cut one side. What good looks like: A clean division of labour between people and AI is where the real time savings live. Revisit it as the AI improves; work that needed a human check last quarter may not need it now.
Workflow has an owner
- 1Nobody owns
- 2Owner unclear
- 3Loosely owned
- 4Clear owner
- 5Owner, actively improving
At the low end: An AI workflow with no owner is one that was set up once and left to rot. Name someone responsible for how AI fits the team's work, with the remit to change it. What good looks like: An owner actively improving the workflow is what keeps AI value growing instead of decaying. Keep that ownership funded with real time; a workflow owner with no hours is a title, not a role.
Time savings measured
- 1Never measured
- 2Anecdotes only
- 3Rough estimates
- 4Measured somewhere
- 5Measured, acted on
At the low end: If you have never measured the time AI saves, you cannot tell a real win from wishful thinking. Pick one workflow and time it with and without the AI; the answer will tell you where to invest. What good looks like: Measured time savings that you act on is how AI stops being an act of faith and becomes a managed investment. Keep measuring as the tools change; last year's win can quietly become this year's wash.