What to Prepare Before an AI Workshop So You Do Not Waste It

The people who move fastest in an AI workshop are not the most technical ones in the room.

They are the ones who walk in carrying one specific piece of work that annoys them.

I have run a lot of AI sessions, from AI Circle classes to training inside companies. The pattern repeats. In the first hour, everyone looks the same: engaged, taking notes. By the third hour, the room has split in two. Some people leave with a tool that actually runs in their job. Others leave with a folder of screenshots they will never open again.

The difference is not intelligence. It is what they prepared before walking in.

Why showing up empty is expensive

An AI workshop is a dense format. Four hours, sometimes two. Inside that you have demos, just enough theory, and hands-on practice.

If you arrive with nothing, the practice gets filled with my examples. You write a prompt for a made-up case, the output looks decent, you nod, you go home. The next morning you face your real work and cannot see the connection.

I know this problem from the other side. At Tokopedia and Flip, I sat in plenty of rooms full of smart people that ended without a decision, because nobody brought concrete context. The conversation floated at the level of "we should be more efficient." The meetings that moved anything were the ones where a single person brought a number and said "here is where it leaks."

Workshops work the same way. The session that changes how you work is the one you fill with your own work.

Three things worth preparing

Nothing complicated. These three are enough, and they take an hour at most.

One: three repetitive tasks you hate.

Not big projects. The small, frequent ones. Writing the Monday meeting recap. Copying data from email into a spreadsheet. Sending replies that are nearly identical every time. Write down three, with a rough estimate of how many hours each one eats per week.

This is your fuel during the hands-on portion. It is also the fastest way to find out where AI actually helps, because repetitive work is the first place results show up.

Two: one example of output you consider good.

If you want help writing reports, bring your best report. If you want help building presentations, bring a deck you are proud of. One file is enough.

This is the step people skip most often, and it matters most. AI is bad at "make it good," because it has no idea what good looks like in your head. Give it one example and quality jumps, with no technique to learn.

Three: access you have already tested.

Your login for the AI tool works. The laptop you bring is your actual work laptop, not a borrowed one. If your company blocks certain tools, ask IT before the day, not in the middle of the session.

This sounds trivial until you price it. Twenty minutes lost to people who cannot log in is twenty minutes nobody gets back.

What you do not need to prepare

This is the part that makes people hesitate for no reason.

You do not need to learn prompt engineering beforehand. That is the material. Arriving half-informed usually anchors you to habits that are not working.

You do not need to understand how the models work internally. I am not an engineer either. What I understand is where work is wasteful and how to point a tool at the waste.

And you do not need the expectation that you leave with everything automated. One job whose time you genuinely cut is a good result for one session. Ten half-built tools is not.

The habit that makes a session feel wasted

Taking notes on everything, trying nothing.

I understand where it comes from. Notes feel productive and carry no risk. Trying feels awkward, especially when your output is bad in front of other people.

But notes are what usually end up as garbage, because what you capture is the sequence of steps visible on my screen, not the judgment in my head about when to trust an output and when to be suspicious. That judgment only sticks when you are the one typing, you are the one who gets a bad result, and you are the one who fixes it.

So if you can only pick one, pick trying. You can write the notes later, from your own attempts, and they will be far more useful.

Why I keep pushing this

Making companies cheaper to run is an old habit of mine, long before AI was a trend: over $4 million per year in savings across a mix of initiatives, from process fixes to cost optimization to product decisions that removed waste.

Back then, that kind of leverage needed a title, an engineering team, and expensive systems. I could find the leak and quantify it, but execution had to queue in a backlog behind dozens of other priorities.

Now AI is the sharpest tool for that same habit, and it can be taught to everyone on a team. My own work is the proof. An 8-channel content distribution system I built and operate alone, with no content team. An Applied-AI Certification system that runs from assessment through to badge. The AI Circle community I run while still leading four product teams.

I am not an engineer. All I have is the habit of hunting for wasted work, and now the tools are in everyone's hands.

That is why your preparation matters. A workshop can only teach you how to hold the tool. The waste you want to cut is something only you can see.

Do it now

Open your notes app. Write down the three tasks you dread most this week. Find one file that represents your best work. Check that your login works.

Fifteen minutes. That is the difference between leaving with a tool and leaving with notes.

If you want to learn alongside people who are testing things and showing each other results, the regular sessions and material live at AI Circle. If the thing you want to move is your team, and you want everyone building the same habit, the format is different and can be fitted to how your company actually works: details on the corporate page.

Bring your own work. That is the only requirement.