Your AI Strategy Is a Gym Membership You Are Not Using.

We have run Weds.ai more than 130 times.

Every Wednesday.
Cameras on.
Hands on keyboards.
52 weeks a year.
We build.

That is the part people do not want to hear when they ask which AI tools we use. I can hand you the list. It will not help you. The tool is not the muscle. The practice is the muscle.

Part one of this series was the Skeleton. Your data. The bones.

Part two was the Circulatory System. Your context. Your second brain.

This is part three. The Muscles. The tools. The part everybody wanted to start with.

Nobody gets strong by buying a gym membership. You have watched somebody sign up in January and disappear by February. Showing up to the gym and doing the work is how you got stronger.

Companies do the same thing with AI. They buy the licenses. They send the all-hands email. They put a slide in the board deck that says AI Transformation. Then they cannot figure out why nothing changed.

They bought the membership. Nobody actually showed up at five in the morning to do the hard work. That is why you are not seeing results.

AI is a muscle, not a project. There is no finish line. You load it, it gets stronger. You stop, it shrinks.

You would not train for one marathon and call yourself a runner.

So which muscles do you build first? This is the Slowinski Pyramid. Three tiers, and you climb them in order.

Get rid of the repetitive tasks you hate doing. Biggest and fastest wins, and almost nobody starts here because it is not exciting.

Then let AI do the things you are not good at. The skills your team does not have.

Then go after the urgent. The thing you need right now, where AI turns a week into an hour.

You cannot start at the top. The organization has to earn its way up.

INSIDE A WEDNESDAY

People ask what actually happens in those four hours. Here is the real thing.

It starts before Wednesday. We keep a channel where we post what is new in AI all week. Almost all of it comes off X. If you are not on X, you are missing the conversation. That is where this breaks first, days before it shows up anywhere else. We built a custom list inside the company so everybody follows the same people and nobody has to go hunting. I will write up who is worth following in a future issue.

Hour one is the discussion. Somebody says, did you see what Boris at Anthropic posted about the five steps of AI adoption? And we talk about it. For real.

I am not standing at the front of the room being a professor. It is a conversation and I am in it.

Do I have to course correct? All the time. Somebody brings a point of view that does not line up with our strategy, and we work it out right there. When the team is aligned, I let them run and we move on. When they are not, we stop and figure out why. That is the job.

The next two hours, everybody disappears and builds. Real work, ideally on what we just talked about. Finance is building. Legal is building. Support is building. This is not an engineering meeting.

The last hour is demos.

Somebody just showed us a way to organize data sources. We are getting buried by Google, chat, email, and our own second brains. He built a system that pulls all of it into context his AI can actually use. Now the whole team knows how it works and we are making it better together.

No consultant. No curriculum. No certificate at the end.

WHEN I KILLED IT

We have run this since November of 2023. Once, we stopped.

In August of 2025 I shut it down. On purpose.

The energy in the room was dying. People showed up, but I could not get anything out of them. Nobody wanted to talk. I could tell they were multitasking, half in the meeting and half somewhere else.

So I said it out loud. This is a waste of time. We are taking a break, and we will come back to it when it is exciting again.

Two months later we did, and the room was different.

Stopping on purpose is a decision. Letting it slide because you got busy is a habit dying. Those are not the same thing.

Now think about the tendons from part one. A tendon ties muscle to bone. Without it, the muscle fires and nothing moves.

That is a disconnected AI tool. It cannot see your customer. It cannot reach your data. So nothing happens.

A tool has to reach two things. Your bones, so it knows what a customer is. Your blood, so it knows how you think. If you built a second brain after the last issue, this is where it starts paying you.

Here is what almost nobody plans for. Muscle regenerates. Your body tears it down and rebuilds it your whole life. Nothing about it is permanent.

Your tools work the same way. The model you run today will be old news in a few months. The tool you love right now will get beaten by something better, probably before you finish rolling it out.

So do not wire a tool so deep into everything you own that replacing it means a rebuild. Keep the bones and the blood in your own hands. Your data. Your context. Those are yours. Let the tools come and go.

YOU SET THE CEILING

Your team will never out-AI you. If you are not in the tools, nobody under you has permission to be. They watch what you do, not what you say. My team does not wait for me to tell them what to build. They watched me build.

That is why I do not skip Wednesdays. Not because I am the best builder in the room. I am not. Because the second I skip, everybody learns this is optional.

Here is what the reps bought us.

AI handles 80 percent of our support. Our AI renewals team converts 14.7 percent better than our human team. I saw that number last week and we made the call on it. Starting January of 2027, every renewal runs on AI. We will use the next six months to move the cohorts over one at a time.

None of that came from a purchase order. All of it came from Wednesdays.

And watch for this one. The better your tools get, the more tempting it is to stop touching them. Let the team run it. Go back to approving instead of building. You will not feel it for six months. Then a competitor does something you did not know was possible.

Here is your figure it out moment. Open your calendar for the next month. Find the block where you personally sit down and build something with AI. Not review it. Not approve it. Not delegate it. Build it.

If that block does not exist, you do not have a tool problem. You have a training problem, and the next thing you buy will not fix it.

So when does it go on the calendar?

Kathy
THE AI BOSS