What to bring to your first AI coaching session

The short answer is nothing.

There is a short intake form before the session and that is genuinely all the preparation there is. You do not need to read anything, install anything, or have an opinion about which chatbot is best. People arrive worried that they are the least prepared person we have worked with. Most people have.

But the hour goes further if a few things are open on your screen when we start. Here they are, in order of how much difference they make.

One: a real week, not a job description.

The most useful thing you can bring is an honest picture of where your time actually went last week. Not your responsibilities, not your title. Where the hours went. Your calendar is usually enough.

Job descriptions describe what a role is for. They are useless for this, because the tasks eating your week are almost never the ones in the description. They are the status update nobody reads, the third time you reformatted the same numbers, the forty minutes of deciding how to phrase one email. Those are the ones worth attacking, and they only show up if we look at the real week.

Two: the task you would be slightly embarrassed to admit is manual.

Everybody has one. Copying rows between two systems by hand. Retyping a report each month because the export is wrong. Reading through a long thread to work out what was decided, again.

These are embarrassing precisely because they are obviously mechanical, and they are usually the fastest wins available. There is no judgment here. Most of these exist because nobody ever had a spare hour to fix them, which is exactly the hour you are booking.

Three: whatever you already tried.

If you have already attempted something with AI and it did not work, bring it. The failed attempt is more useful to us than a clean start, because it tells us what you expected, what the tool actually did, and where the gap was. Often the idea was right and the approach was wrong by a small amount.

If you have a subscription you are paying for and barely using, say so. That is a common place to start, and it is cheaper than adding another one.

Four: your constraints.

Anything you are not allowed to do, we should know at the start rather than after we build something. Data that cannot leave your organization. A tool your employer blocks. A system you are stuck with. A regulator, a client contract, a policy.

What we do with all of it

The first hour is mostly sorting. We go through your week and put things into three piles: work AI can carry, work that still needs your judgment, and work that should not be happening at all. That third pile is often the largest and the most satisfying.

Then we pick one thing. Not the biggest thing, the one with the best ratio of time saved to effort required. And we build it, in your accounts, on your real work, while you watch and do most of the typing. You should be able to redo it without us the following morning. That is the test.

What you leave with

A written roadmap of where AI realistically saves you time in your particular job, ordered by what is worth doing first, and one workflow already built and running. Not a slide deck, not a reading list, not a set of prompts to try later.

And a more useful thing that is harder to put on an invoice: a working sense of what these tools are actually good at. Once you have that, you stop asking whether AI can help with something and start recognizing the shape of the problems where it can. That recognition is the skill. Everything else is details you can look up.

What this is not

It is not a course. There is no curriculum and no homework. It is not a tool demo, and we do not sell software or take anything from the companies whose products we might suggest. And it is not a pitch: if an hour with us is not worth your money, we will tell you inside that hour rather than after it.

Afterward

Some people take one session, build the thing, and never come back. That is a fine outcome and we would rather have that than a package you felt obliged to buy. Others go weekly for a while, then monthly, then stop. You set the cadence, and you can stop at any point.

The only thing we ask is that you actually use the workflow we built. A session that becomes a note in a folder helped nobody.