My team has been working on AI since November 2023 when ChatGPT came out. Our weekly AI usage hit 82%. That means 82% of the time we spend working, we are using some AI tool.
That sounds like we are ahead.
We are not.
120 Features in 90 Days
Claude shipped over 120 features in the last 90 days. Opus 4.6 with a million token context window. Cowork. Computer use in the browser. Claude Code in the app. Plugins. Connectors. Dispatch. Skills.
That is one product. From one company. In one quarter.
There is no way to keep up with this unless you are spending real, dedicated time. Not reading about it. Not watching a recap video. Sitting down, opening the tools, and building things with them.
My team is doing this every single day and we are still behind. So I want you to really hear me when I say: if we are behind at 82% usage, what does that mean for the team running at 10%? Or 5%? Or the team that tried AI once in January and moved on?
It means the gap is getting wider. Every week.
If your team is not spending dedicated hours every week inside these tools, you are not falling behind slowly. You are falling behind fast. And the teams that are in the tools are compounding their advantage while you sit still.
What We Are Doing About It
One of my team members came to me last week and said: can we bring back the all day hackathons? We need real time to get our hands dirty.
I took it to our monthly all hands. I asked the team: would you be willing to go back to half day Weds.ai sessions? Three hours of deep diving. One hour of demos.
They said yes. Every one of them.
So we are back. Four hours a week. Protected time. Cameras on. Hands on keyboards. No spectators.
We did this before when things were moving fast. Then it calmed down for a while. The tools felt stable. We pulled back to shorter sessions. And then the release velocity picked back up and suddenly we were behind again.
That is the pattern. And if you are a CEO, you need to understand this pattern because it will keep repeating. AI does not move at a steady pace. It surges. And when it surges, you either invest the time or you lose ground you will not get back.
What Enablement Actually Looks Like
Let me be specific about what we are doing because "enablement" can sound like corporate training. This is not that.
Four or five months ago when Claude Code launched, I made sure every single person on the team had it installed and running within two weeks. Terminal access. Set up. Working. Not optional.
Now the questions are different. Do the non-technical people know how to use GitHub? Does everyone know best practices for deploying on Vercel? Are people comfortable finding a repo, making a change, and pushing it to production? Are they getting the most out of features that did not exist eight weeks ago?
The answer was: not yet. So we are fixing it. Person by person. Skill by skill. Wednesday by Wednesday.
This is what it looks like to treat AI as a muscle. You do not work out once and call yourself fit. You train consistently. And when the weight goes up, you adjust your program.
The weight went up. 120 features in 90 days is the weight going up. Our response is not to read about it. It is to get back in the gym.
The Mandate
This quarter, every single human on my team will build something and put it into production for other people to use. Every one of them.
That feels like a high bar. My head of product raised it higher. He wants every person to build a feature or function that makes one of our existing products better. In production. For real customers.
Think about what that means. An account manager takes a call with a customer. The customer asks for something. That account manager finds the repo for that product. Builds the feature. Sends it to a test environment. Gives it to the customer a day or two later.
"I am not technical" is a choice. My SVP of commercial has never written a line of code in her life. This week she built a predictive retention dashboard. You click on a customer name and it tells you, down to the dollar, what happens to revenue if that relationship stays or goes. Not a spreadsheet. Not a Google Sheet someone updates on Fridays. A real tool that anyone on the team can log into and use.
We have wanted this for years. We had back of the napkin math. We had iterations. We never had a single source of truth that showed the actual impact of keeping or losing a customer.
Now we do. Built by someone with zero technical background. In a week.
The CEO Problem
Here is the thing I keep coming back to. Your team will never out-AI you. You set the ceiling. If you are not in the tools, your team has no permission to be.
When I say we are going back to four hours a week, I do not mean my team is going back to four hours a week. I mean I am going back to four hours a week. I am in the room. I am building. I am demoing what I made. And then they do the same.
That is how culture changes. Not with a memo. Not with a mandate from someone who has never opened the tool. With a CEO who sits down, opens the laptop, and builds something in front of the team. Every single week.
52 Wednesdays. Cameras on. That is how culture changes.
We calmed it down for a while. Now we are back. Because the tools moved and we need to move with them.
If your AI usage is lower than 82%, you are behind. If it is 82%, you are still behind. The only question is what you are doing about it.
We are doing four hours a week. We are mandating production builds from every person on the team. We are treating "I am not technical" as a choice, not a limitation.
That is the corrective action. Not a strategy deck. Not a committee. Time in the tools.
If I can do this, so can you.
Kathy