Your AI Intern Needs a Job Description. Not a Pep Talk.
Why most AI initiatives stall, and how the smartest teams are redesigning work instead.
12/27/20252 min read
🎢 The problem no one admits out loud.
Most organizations say they’re “adopting AI.” What they’re actually doing is introducing a very powerful intern… and giving it motivational speeches instead of structure.
You’ve seen it:
AI tools rolled out with excitement
People told to “experiment”
Leaders hopeful productivity will magically appear
Then six months later:
Adoption is patchy
Risk teams are nervous
Employees are confused
Value is hard to prove
The issue isn’t the technology.
It’s that the work was never redesigned.
AI doesn’t fail because it’s not smart enough.
It fails because no one told it what job it’s allowed to do.
💡 A simple idea from Prosci that changes everything.
At a recent Prosci webcast, Integrating AI to Unlock Your Team’s Human Potential, one idea cut through all the noise.
Don’t start with tools.
Start by sorting the work.
Every role, every team, every process, can be broken into three buckets:
🧩 The three buckets that make AI usable:
🧠 1. My Work. Human only.
This is the work that should never be handed to AI.
Judgment calls
Ethics and accountability
Trust-building conversations
Reading the room
Making trade-offs when there is no perfect answer
AI can inform this work.
But humans own it, full stop.
🤝 2. With Me Work. Human + AI together.
This is where AI shines as a thinking partner.
Drafting content
Summarizing inputs
Analyzing patterns
Generating options
Stress-testing scenarios
The human leads.
AI accelerates.
This is where speed increases without sacrificing control.
⚙️ 3. For Me Work. AI automation.
This is the work people quietly hate.
Status updates
Data pulls
Meeting scheduling
Basic reporting
Reminders and follow-ups
It’s repeatable. rules-based. draining.
This work should move out of human hands.
🔄 Why this framing works so well.
This simple map does four powerful things at once:
Reduces fear
People can see what stays human, nothing is being “taken away” blindly.Improves speed
Teams stop debating tools and start redesigning workflows.Makes governance practical
Risk and compliance can point to exact steps where AI helps and where humans sign off.Turns adoption into something measurable
Every activity has an owner, a reason, and a success signal.
This is how AI moves from “interesting pilot” to operating model change.
🏢 What this looks like in real transformation work.
Let’s ground this in something familiar. change and transformation.
My Work.
Building trust with leaders
Navigating resistance
Aligning sponsors
Making sense of competing priorities
With Me Work.
Analyzing survey and listening data
Drafting communications and talk tracks
Synthesizing feedback themes
Designing learning journeys faster
For Me Work.
Survey administration
Status reporting
Meeting coordination
Routine progress summaries
AI doesn’t replace the human side of change.
It protects it.
⏱️ Try this on Monday. seriously.
You don’t need a roadmap or a new platform to start.
Take 15 minutes.
List 10 activities your team does repeatedly
Label each one.
My Work
With Me
For Me
Add one sentence explaining why it belongs there
Pick one activity to move this month
That’s it.
You now have an AI adoption plan that is actually about work.
🚀 The real takeaway.
AI isn’t here to replace humans.
It’s here to force clarity about where humans matter most.
When leaders stop giving AI pep talks and start giving it clear jobs, something powerful happens:
People trust it
Risk teams relax
Adoption sticks
Value shows up
That’s when AI stops being a feature and starts becoming part of the workforce.
🤔 One question to leave you with.
If you mapped your team’s work today:
What would you automate immediately to give people time back?
And what would you protect as human-only on purpose, no matter how good AI gets?
That conversation is where real change begins.
🧪 Kenny’s Change Lab
Where AI, work design, and human potential meet.