There’s a scene from Office Space I keep coming back to in client conversations. It’s when the Bobs (a pair of consultants who have been tearing through a company’s org charts to recommend layoffs) turn the tables on a manager who’s only been asking his team about the cover letters on their TPS Reports.
If all you’ve been doing is working on TPS reports your entire career, and suddenly TPS reports stop mattering to the business, do you still have a job? Probably not.
In many organizations, somebody designed a role that asked a human being to perform a rote task, year after year, and called it a career. That’s an organizational failure as much as an individual one. And AI is exposing it everywhere.
How AI surfaces dysfunction you already had
When our team helps a client build the knowledge foundation that makes AI actually useful — the documented context, the playbooks, the shared understanding of how work happens — something uncomfortable surfaces almost every time. Leaders discover they never knew what people were actually spending time on. That knowledge was squirreled away in very specific silos.
On one recent software project, the client hadn’t revisited their original business requirements in years. Nobody could explain why certain processes existed. Legacy commitments were eating resources nobody had mapped. Documentation gaps existed because someone, somewhere, felt nervous or overprotective about what they knew.
None of that dysfunction was created by AI. It was sitting there the whole time, invisible and holding weight, like a cracked foundation you only notice when you finally pull up the floorboards. AI just made it visible faster.
So when a role turns out to be fully replaceable by an automated workflow, the honest question is rarely about the person in the role. The honest question is why the organization designed that role to require so little thinking in the first place.
Two leaders, same technology
In some organizations, AI adoption results in a very cold process of layoffs. In others, it results in people getting a bigger voice in what they’re doing — sharpening their skills, contributing closer to the core of the mission. Same technology. Same economic pressure. Completely different outcomes.
The difference comes down to something that the Bobs actually deliver as a punch line. It’s about being the kind of leader who genuinely wants a team full of excited, capable, engaged professionals, versus the kind who wants to check the box and get to their tee time on time.
The check-the-box leader looks at AI and sees a headcount reduction they can call “transformation.” They cut the people, keep the broken processes, and hand the remaining team a chatbot. Then they’re shocked when quality collapses, because the people they removed were the ones quietly shoring up the workflow the whole time.
The leaders getting this right are doing something harder. They’re using AI to strip the rote work out of roles so the humans in those roles can finally do more of the thinking the organization needed all along. They’re redesigning the operating model instead of just turning off the lights in half the building.
What AI can’t do
AI cannot lead a team in a way that makes people feel like they’re contributing at a higher level. Your “Chief of Staff” agent is never going to become an inspirational leader. On a good day, it can find a better way for your team to get their timesheets filled out correctly. On its best day, it can design an automation to eliminate a timesuck routine.
I spent the first post in this series on why: large language models are built to give you the answer you seem to want, and a team member who pushes back with enough context to explain why is doing something the model can’t. That limit is structural. No future release fixes it.
When people on your team feel like what they do can easily be replaced by an automated workflow, they need a genuine opportunity for more agency — with the product, with the service, with the people your organization serves. Without that, you end up asking whether the work is even worth doing. And your team asks that question long before you do.
In smaller organizations, your differentiator was never going to be your technology stack; any competitor can buy the same tools. Human creativity is what separates you from the field, and it’s what gets a hesitant supporter or client to move at all. Automation alone never has.
Your people will also determine how your organization treats the communities you serve. Teams that feel replaceable tend to serve people like they’re replaceable too. That’s a slow, quiet way to drift from your mission — and it’s one of the harder things to audit from the inside.
Bigger voice or smaller role
Every AI decision you make right now answers one question, whether you intend it to or not: am I giving my people a bigger voice, or a smaller role?
A bigger voice looks like this…
The junior analyst who used to compile status reports now interprets them and pushes back on the plan. The coordinator who used to copy data between systems now redesigns the intake process because she finally has time to notice it’s broken. The person who ran the rote workflow becomes the person who improves it.
A smaller role looks like quiet scope shrinkage.
Same title, less trust, less judgment required, and a growing sense that the automation is the real employee and the human is the backup plan. People feel that shift immediately. They start protecting knowledge instead of sharing it — which, in a moment when your AI systems depend entirely on documented, shared knowledge, is exactly the behavior that will sink your implementation.
And AI work runs in loops. Every cycle of iteration needs humans with enough context and enough agency to say “this output is wrong” and explain why. Gut that capacity from your team and you’ve automated your way into a building with no inspectors.
The uncomfortable audit
If you lead a mission-driven organization and you want to use this moment to redesign rather than just reduce, start here.
Map what your people actually spend time on — the real work, including the invisible commitments nobody has documented. Identify every role that’s been designed around a rote task, and treat each one as a design flaw you inherited rather than a person you need to remove. Then sit down with the people in those roles and ask what higher-level contribution they could make that would actually show up for the people your organization serves.
Do that before you sign a single AI contract. The technology will expose your organizational design either way. You get to choose whether it exposes a leader who built something better, or one who checked the box.
If you want a grounded place to begin, start with an honest outside read of what your audience actually experiences today. Our Website Reality Check exists for exactly that first step. Take it, share the findings with your team, and let the redesign start from what’s true.
