Lean In.

AI first is a habit.

The ClickUp post will get flattened into a layoff story.

I get why. When a CEO says the company reduced headcount by 22%, then mentions million-dollar salary bands for people creating outsized results with AI, the obvious reaction is emotional. People will argue about whether it is fair, whether it is cold, whether the 100x framing is useful, and whether any company should talk about people this way.

That conversation matters.

But the part I keep coming back to is smaller and more practical.

What does AI-first mean before it becomes a restructure?

Because that is the bit most people can actually control.

The slogan is too big

“AI-first” has already started to become one of those phrases that gets repeated until it stops meaning anything.

Some people use it to mean every employee should use ChatGPT.

Some use it to mean every workflow should be automated.

Some use it to mean no headcount gets approved unless a manager proves AI cannot do the work.

Some use it as a signal to investors that the company is serious about productivity.

That is the problem with big phrases. They make people nod without changing the work.

If AI-first is only a company value, it sits in a slide deck. If it is only a cost program, people resist it because they can feel the threat. If it is only a tool rollout, adoption peaks in week two and quietly falls back to old habits.

The useful version has to be more boring than that.

AI-first means every task gets questioned before it gets done the old way.

That is it.

Before you write the update, build the report, triage the tickets, review the spreadsheet, prepare the meeting notes, clean the data, draft the Jira ticket, or send the follow-up, you pause for ten seconds and ask:

Could AI improve this?

Could AI do the first pass?

Could AI check my work?

Could this run without me next time?

Could this become a reusable workflow instead of another one-off effort?

That is the muscle.

This is where people miss it

Most AI adoption fails because people treat AI as an extra step.

They do the task the normal way, then ask AI to polish the output.

Write the email, then ask AI to make it cleaner.

Build the deck, then ask AI to rewrite the bullets.

Run the meeting, then ask AI to summarise the transcript.

That can help. But it does not change the system.

The bigger shift happens when AI moves earlier in the work.

Not “make this better after I finish.”

“Help me decide what this should become.”

“Find the pattern before I spend an hour reading.”

“Turn this messy input into a structure I can use.”

“Watch this folder and tell me when something changes.”

“Take this repeated action and make it happen on a schedule.”

That is where the work starts to bend.

I have felt this most clearly in my own systems. My Obsidian vault does not just store notes. It has skills that process clippings, prep content, capture meetings, sync project state, and turn rough input into usable drafts. Some of that runs from a command. Some of it runs on a schedule. The point is not that the setup is perfect. It is not.

The point is that the default question changed.

I do not look at a repeated task and think, “I need to remember to do this every Friday.”

I think, “Why is this still depending on me remembering?”

That is the practical meaning of leaning in.

Leaning in is not blind optimism

I do not think this means handing every task to an agent and hoping for the best.

That is usually how teams create more work for themselves.

AI can generate noise at a speed humans cannot review. It can make weak assumptions sound finished. It can automate the wrong step and make a bad workflow harder to see. It can move faster than your permissions, your data rules, and your internal trust model.

Inside a company, those things matter.

Governance is not a blocker to AI-first work. It is part of the work.

The useful question is not “Can AI do this?”

The useful question is “What part of this should AI do, under what constraints, with what human check?”

That changes the shape of the answer.

For some work, AI should draft and a person should approve.

For some work, AI should classify and a person should handle the exception.

For some work, AI should watch for changes and alert someone.

For some work, AI should do nothing because the human moment is the point.

That last one matters. The ClickUp post mentions that the 100x organisation still depends heavily on people. I think that part gets lost because the compensation headline is louder. But the best AI-first teams will not be the ones that remove humans from every workflow. They will be the ones that know exactly where humans are still the highest-value part of the system.

Customer conversations. Taste. Judgment. Trust. Context. The ability to sense when the output is technically correct but still wrong.

Those do not disappear because a model can write faster.

The job is moving upstream

The work is shifting from doing every step by hand to designing better loops.

That sounds abstract until you make it concrete.

A weekly report used to mean gathering inputs, checking numbers, formatting notes, writing the summary, and sending it around.

AI-first means asking which parts repeat.

Can the inputs be pulled automatically?

Can the first summary be generated?

Can the anomalies be flagged?

Can the report draft land in a place where a human reviews only the judgment calls?

Can the follow-up actions become tickets without someone copying and pasting them?

Now the job is different. You are not just writing the report. You are designing the report system.

That is the career move people should pay attention to.

The high-value person in an AI-first company is not the person who knows the most prompts. It is the person who can look at messy work and see the system underneath it.

Where is the repeatable pattern?

Where does context enter?

Where does approval matter?

Where does quality break?

Where does the work depend on one person remembering?

Where could a small agent, script, or workflow remove the drag permanently?

That is a different kind of skill. It is part operator, part builder, part product thinker.

You do not need to become a software engineer to build that muscle. But you do need to stop treating your current workflow as fixed.

Start smaller than the headline

The ClickUp post talks about 100x organisations and million-dollar salary bands.

Most people do not need to start there.

Start with the task you hate repeating.

The thing you do every week that always feels slightly beneath your actual ability.

The update you rewrite from scratch even though the structure barely changes.

The folder you clean manually.

The ticket you shape from the same rough notes every time.

The meeting notes you turn into actions.

The spreadsheet you scan for the same three issues.

Take one of those and ask what AI could do before you touch it.

Then ask what would need to be true for that task to run again next week with less input from you.

That second question is the important one.

Using AI once is help.

Changing the way the task runs is where the gains compound.

There is a reason I keep coming back to boring workflows. Boring workflows are where the compounding sits. Nobody transforms their work by asking AI to rewrite one email. They change their work by removing the small repeated drains that quietly consume attention every week.

That is how you create room for better work.

Not by waiting for the company to announce an AI-first strategy.

By making your own work harder to waste.

The uncomfortable part

There is a hard edge to this.

If your job is mostly moving information from one place to another, AI will put pressure on it.

If your value is mostly knowing where the template lives, AI will put pressure on it.

If your workflow only works because nobody has had time to question it, AI will put pressure on it.

I do not think pretending otherwise helps anyone.

But the answer is not panic. It is to move closer to the system.

Understand the work well enough to redesign it.

Understand the tools well enough to test what is possible.

Understand the risks well enough to know where the human check belongs.

Understand the business well enough to automate the right thing, not just the visible thing.

That is what leaning in means to me.

Not cheerleading every AI announcement.

Not accepting every executive narrative about productivity.

Not turning work into a race against machines.

It means building the reflex before someone else builds the restructure around you.

Every task is a chance to ask whether the old way still deserves to exist.

That question is going to decide more careers than the tool choice.